# The story of Anima Anandkumar, a learning machine guru with Amazon AI powers

Cooper, the father of Visual Basic Family Engineer Anima was born in Mysore to raise his parents. His father was a mechanical engineer (IIT Madras) and his mother was an electronics engineer. His grandfather was a math teacher. Anima naturally enjoys science and mathematics. She says, I think mathematics is a natural language for talking and understanding the world. Anima with his family in front of the Mysore Palace. Anima visited Marimallappa High School in Mysore, where the school’s principal, MN Krishnaswamy, a math teacher, also taught students the philosophy of life. Anima learned why we solve some problems rather than memorizing them blindly. As a child, she was always amazed to see how science led us to the moon and beyond. She said: For me, it’s a miracle, but at the same time, something we can understand and deepen. ?? Anima loved to dance Bharatnatyam and started learning it when she was three years old. Animaare’s father is a businessman and he brought her to the industry to show how the machine works. She misses many trades when she was a child. She also danced at the age of three. She remembers, My mother took me to the dance class just to see but I started learning from there. I have been doing Bharatnatyam for many years and doing tests there. At school, I danced and choreographed. I always thought that if I did not put an end to technology, I would become a dancer. Anima ranked fourth in the state on its 10th grade exam. You have heard of his father’s IIT-JEE. Thanks to Anima, IITs are the pinnacle of good mathematics and science. Since there is no training institute for IIT-JEE in Mysore, she did distance learning, called academic books (like Feynman’s lectures on physics). She said, I cherish these times because I can understand specifically how to take this test as a challenge. A lot of thought and inner spirit helped me do that. Life at IIT Madras Anima joins IIT Madras and elects electrical engineering. She is part of the organizing group of the cultural and technical festival of the school. But the highlight of his college years was co-founder of the Robot Club at IIT Madras. In her last year at university, she and her teammate expressed the idea of ??the Robotics Club to the director. The club not only helps students to relax and unwind, but also gives them more technical knowledge. Anima at IIT Madras hostel Anima believes that IIT students have a soul business. She recalled an incident when they met a villager who built a very handy device to remove the coconut with minimal effort. She and her friends helped her market it. After the third year, Anima spent the summer at IISc under the direction of Professor YB Venkatesh where she was introduced to the Gabor filter. This helped in his research in image processing (multi-effects resolution). That’s when she decided to pursue her PhD. She has applied at universities across the United States and has received a call from Cornell University (who eventually became her mentor). Animaare B.Tech. The project is based on the idea of ??using segmented pollution transformations to efficiently process images. It belongs to the electrical engineering department. It was the beginning for her to think about statistical signal processing and machine learning. Anima in his congregation at IIT Madras Scattered Discovery Network, also known as the Internet of Things Anima was a bit scared at first. She said, So far, you have to solve a problem where you have a question. But now you have to ask a question. Anima graduated from Lang Tong Ph.D. During the first week, Ms. Tong’s mentor gave her a book on physiology Performance Noga Alon, and asked him to read it and present That’s what she did three days later. Anima thinks that many researchers do not spend enough time thinking about what we need to solve. After the first year, she tested the field of distributed sensor networks, now called “Internet of Things.” She has worked on resolving issues such as “how do we connect multiple devices when they have limited battery and communication capabilities, but all have a common goal.” < With network sensors, we have sought to provide algorithms that can be implemented with lightweight communication mechanisms and can communicate . < It is important to introduce This is one of his first works, in which he suggests that the bandwidth requirements (for transmitting sensor data) can be significantly reduced by Statistical inference instead of sending raw data. Later, she wrote an article extending more about this work and called communication based on the type. > The idea is that the sensors that should send their accounts see how many entries of each type instead of the actual entries. This allows for better and more efficient bandwidth savings as they do not. Decode the raw data of each sensor. Anima won the award for best author of Anandkumar and L. Tong, “random access type to detect distribution on fading multiaccess TV channels” IEEE Ceiling . Signal Proc., Vol. 55, pp. The problem of accessing the opportunity spectrum, where she tried to answer the question, “Can we transmit information more effectively by licensed players? We can have the opportunity for actors When the main user does not use the tape, this is called cognitive radio access. Anima has built up the probing problem – which groups have the freedom to (statistics) and exploit them. This is never going to take off but the idea is to persuade agencies to also allow universal users to transmit in the frequency bands as long as they can prove that they will. do not interfere with ng Reading related – Meet the co-founder of the Julia programming language, Viral Shah IBM There are a lot of transactions going through a legacy system where we lose track of what each side is doing, its configuration, and what it is supposed to do. Anima has been working to resolve this transaction-level transaction tracking problem. Old systems do not have built-in device support. Anima designed the framework to model these transactions through a state transition model. They will also put dot time stamps in different positions to track them from start to finish. His team has filed a patent on how we can selectively refine the existing system. At the end of her doctorate, Anima finds herself at the fell private industry and universities. She always enjoys math and wants something that is based on theory, but that it can itself take all the way to practice. In the last year of his PhD, she contacted Alan Willsky, a professor at the Massachusetts Institute of Technology. She is Anima Anandkumar pioneered the research of finding global optimal in non-convex problems, a big pain point in machine learning. Our protagonist for this weekare Techie Tuesdays, Anima is an academician who represents the best of both worldsindustry and academia. She has contributed significantly to major AI and ML projects at Amazon.This is a treat for all machine learning enthusiasts. In my two hours of conversation with Anima Anandkumar, Principal Scientist at Amazon Web Services, I was injected with the most potent dose of technical knowledge. Not that I didnn’t expect it while talking to an ex-faculty of UC Irvine (soon to be an endowed professor at Caltech), known for her research on non-convex problems (in deep learning).Our Techie Tuesdays protagonist of the week, Anima has worked towards establishing a strong collaboration between academia and industry. She follows an unconventional style of teaching, the one she would have loved as a student. Her love for mathematics and specifically for tensor algebra has made her solve the toughest of the problems in machine learning with much ease.Hereare the story of a machine learning guru who could have also been a Bharatnatyam dancer.Also read – The untold story of Alan introduced to the idea of probabilistic graphical models which is about modeling dependencies between different variables in the data, i.e. how is the data collected in one sensor related to the data collected in the other. These correlations could be efficiently captured through graphs (dependency graphs). That’s when Anima switched to machine learning because she wanted to deal with the core problem of modeling lots of data efficiently. Anima with her postdoc advisor Prof. Alan Willsky and her fatherShe wanted to design learning algorithms that can process at scale and make efficient inferences about the underlying hidden information. She recalls,That was the start when I felt like I could also program these algorithms, try it out on real data and be able to see the tangible results which showed how beautifully the theory worked. Anima got an award at UCI for research excellenceShe found a whole range of applications where she could apply thiscomputer vision, time series analysis (like stock market), time-varying data. She interviewed with many companies while graduating with her PhD and even received some offers. But she accepted the offer from UC Irvine to be a faculty because it allowed her to do both research and work in the industry. She wanted to advise students and teach in her own way to encourage more students to take up sciences.Research at UC IrvineWhen Anima joined UC Irvine as a faculty, that time was the beginning of the big data revolution. People were trying to figure out what were the algorithms that would allow us to process data at scale and also make useful inferences. She started working on the problem of latent variables. These are the hidden variables which are always hidden to us when we make any number of observations. She says, There are some hidden aspects that we may never be able to identify because we just don’t have enough information. But the ones which are limited by the computing capabilities can be found.She worked on hierarchical categorisation which includes identifying the hidden tree that related the different variables. For example, in image classification of a house, there can be indoor, outdoor and bedroom, drawing room, and bed, sofa and so on. The idea is to have the algorithms detect these classifications from data. Anima came up with efficient, fast algorithms that enabled to do this practically well and in theory established that this model is guaranteed to give you a right answer under a certain set of conditions.Of tensors and tensor algebraTensors and tensor algebra give rise to a lot of interesting mathematical properties. With tensors, the notion of multi-linearity gives rise to much richer operations. This includes hidden phenomenon like learning about hidden communities in social networks, i.e. figure out the common interests between two people who’re friends on a social network and then use it to suggest friends. It enables to process large graphs with millions of nodes within an hour using these algorithms.Anima says,Kruskal who gave the famous algorithm on minimum spanning tree which is one of the classical algorithms in computer science, also gave elegant results on what can you identify using tensors (called the uniqueness theorem). Since then, there’s a treasure trove of problems. This has deep connection with deep learning on how we can resolve non-convex problems at scale.Solving the non-convex problem Aniim