Even the most powerful supercomputer in the world falls short when it comes to matching the raw computational power, efficiency and energy consumption of the human brain! However, one can always take inspiration from nature and make an effort to solve one or many of the existing problems in the world.
Pratik Kumar, our next pathbreaker, PhD candidate at IISc Bangalore, takes a multi-disciplinary approach to achieve the end goal of designing highly efficient computational architectures that are biologically inspired by how human brains work .
Pratik talks to Shyam Krishnamurthy from The Interview Portal about taking inspiration from the fact that vey little is known about the human brain and attempting to build Integrated circuits that can do cognitive tasks like the human brain, which could potentially help in bridging the gap between humans and machines.
For students, some fields of research are so niche and unexplored that there is no information to guide you. But you can also see this as an opportunity to learn and advance technology in areas that the world isn’t aware of !
Pratik, tell us about Your background?
I grew up in the capital city, Patna of state Bihar. As a kid, I was very active in participating in various science experiments, presenting projects at different states and national events, trying out new things and finding ways to do things that were different from society’s standards, and indulging in a lot of extra-curricular activities. Both of my parents guided me in my studies till I crossed my 8th standard. After 8th, I started forming my own effective ways to study better by completing the upcoming syllabus in advance, starting preparations for the next semester beforehand and finding new ways to learn things quickly. The sole reason was that I started loving my studies and formed a different perspective towards it. I understood that education was not merely to achieve grades but a way to achieve something more significant. I also understood the sacrifices parents make to educate their children. However, I initially didn’t plan anything regarding my research career. I wasn’t even aware of IITs and NITs till I completed my secondary school examinations. In senior secondary, I was introduced to all those exciting fields, and branches of science and maths and simultaneously learned the extent of competition out there. After my senior secondary, I realized how important it was to plan ahead for your future, at least for the upcoming few years. I had all the support from my family to choose whatever career path I wanted, and to date, I still involve them in every step of my decision, but it is always entirely up to me to decide what direction I want to choose; whether it is an engineering field, research field or job.
My father is an advocate, and my mother is a housewife. My father had struggled a lot to get his education and reach where he is today, as he had to start everything from scratch since the previous generations of my family were not much educated. Not having the support from my grandfather, he had to earn for his education, and I could only imagine how difficult it would have been at that time. That’s why both my parents understand the importance of education and what kind of changes education can bring in one’s life. I believe it’s an advantage of having been born in an educated family. I also have a younger sibling who is pursuing law at National Law University.
I do enjoy various sports, be it physical or mental. However, I never stick to it because of the tight schedule I have, and of course, the other reason is that once I learn something, it becomes boring. So I generally love finding tasks where I can keep on learning new things. This also became one of the important reasons why I joined as a researcher in IISc .
What did you do for graduation/post-graduation?
I had quite a bit of interest in electronics from my senior secondary itself. Though it was not clear what electronics is at that time, I guess as a kid I was just fascinated and curious after seeing multiple components inside cell-phones, televisions, etc. I just wanted to make sense of it somehow. I guess it shaped my initial interests in the branch which I chose later on in my life.
After my senior secondary and after many challenges in selecting a college, as I couldn’t score well in IIT-JEE and other national competitive examinations, I finally decided to pursue my bachelor of technology in Electronics Communication Engineering from Amity University. I decided to give everything I had in me and not lose out on this opportunity and waste even a day. I immersed myself in studies as I was aware of what was at stake at that time. I started interacting with faculty who had vast experience and formed a great team of friends with whom I could talk and interact. My learning journey started from my bachelor’s degree onwards.
There I met a few great professors whose guidance and perspective changed my perception and thinking. They taught me how to think keeping the long-term in mind rather than just focusing just on a step ahead. The overall result of that dedicated preparation was that it cleared up my fundamentals to a great extent. By the time I completed B.tech, I already had 5+ job offers in hand while still preparing for competitive exams. I decided to join a reputed MNC. I worked for two years in a reputed MNC as a system engineer, but quickly realized that I was not putting my knowledge to the best use of my capabilities.
With nothing new to learn there, I soon realized that I was not making good use of my knowledge, so I decided to give some competitive exams and join a core institute for my higher studies. Hence I pursued my masters from the Indian Institute of Technology to further expand my knowledge in my field. I joined the Indian Institute of Technology in my branch of interest (VLSI) where I had the privilege of completing my masters’ project under a well-experienced professor. During my master’s, I had developed interest in unexplored areas of cutting-edge research across the world and its impact on human life and society, after which, I decided to pursue a career as a researcher.
One gets to learn a lot when we work under such a professor, since we get the benefit of all those years of experiences and knowledge they have, but one thing was clear. For myself to get that knowledge and experience, I had to put in a lot of effort since it was my responsibility, not theirs. During the final year of my master’s, I was exposed to the field of neuromorphic and brain-inspired computing. This niche field was still in a nascent stage in India but had tremendous potential in the future. So it was also challenging to find an excellent institution to do cutting-edge research in this field. After a few months of unending interviews in the older IITs and IISc, I finally found a young and energetic professor in IISC who exactly matched my criteria. I decided to carry on with further research in the field of Brain-Inspired Artificial Intelligence, Brain-Inspired Computing, and Neuromorphic computing. All these fields are interrelated in one way or another.
So I joined IISC as a researcher.
Furthermore, research (PhD) is not like a masters’ or bachelors’ degree where you get a degree in a fixed amount of time. A research degree is always earned rather than being given like a masters/bachelors degree. Research is a long-term commitment. So another challenge in front of me was choosing a professor who had experience in my field of interest. My selection criteria for the research was based on, how good the professor is, how much time he/she spent with her students, who was his supervisor, will i have enough freedom to work in the lab as per my own choice, is there a time restriction (like it shouldn’t be a 9 to 5 typical corporate job) and believe me, I was fortunate enough to be selected in one such lab. It is one of a kind academic research lab in India that focuses on multiple fields and is research-oriented, keeping the final product outcome in mind. And to add to it, it is always a boon to work with a young professor who thinks out of the box. I finally decided to work with a young professor in IISC who exactly matched my criteria. Just to add, I was selected for my doctoral degree in almost all the Older IITs wherever I applied (7 IN TOTAL) as well as in IISc. Instead of just being a Ph.D. degree, I wanted it to be a more research-oriented degree. In a Ph.D. one should always remember that we fail multiple times but succeed only a few times, but those numerous failures are necessary for the few final successes.
What made you choose such an offbeat, unconventional and uncommon career?
I pursued my bachelor’s in electronics and communication and my master’s in one of its specializations (VLSI). I always wanted to enrich my knowledge in the domain of my interest. During my final year of my masters’ degree, when I was working on a project under a renowned supervisor, I developed an interest in cutting-edge research that could change the way we do computations today.
I still remember that I just completed the initial year of my masters’ in IIT and was looking for a research project where I could apply/test all the knowledge I gained in the VLSI field from my graduation until the end of 1st year of my masters. Being the topper of the class, it was entirely up to me to decide under which professor I wanted to carry out my master’s project. The selection of the professor here turned out to be a crucial point and was more difficult than expected. I had to ensure a balance in choosing the professor and the project offered by them. For project selection, my criteria was that it should not be in a saturated field and the project should be related to ongoing cutting-edge research as well as aligned with my area of interest. I could have chosen a young professor or an experienced one. Finally, I decided to choose a well-experienced professor whose project/research area matched my area of interest, partly because I wanted to benefit from all those years of experience to make my fundamentals strong and get started with the research. It is truly said, “what experience teaches us can’t be learned in textbooks”. I started my final year master’s project related to CMOS circuit emulation/design of novel multi-state memory elements and their applications. While doing the literature survey for the same, I was introduced to the field of neuromorphic computing and Brain-Inspired computing. I learned how these novel memory devices could help implement energy and area-efficient architectures that mimic human brains (or at least try to). Digging deeper into the same, I realized that this field is very niche and very promising. Since the field was/is very new, the only challenge was that the guidance available to carry out the research was very limited. None of the online courses/open-source materials were available and to my surprise, a minimal number of people in India were just starting to work on it. And, only top universities or research-oriented companies were working on it. On digging further, and looking at older IITs and IISc, I found a few professors working on it. The research had just started in this field at that time in India and from the hardware perspective, it was only in IISc. To my surprise, there was just one IIT working on the same field but on the device side (which was not an area of interest for me). Once I realized this, everything was clear. What was left was to carefully lay out my research plan, fulfill the requirement for selection, get some good publications, and clear the fundamentals of research where my final year supervisors’ experience and knowledge helped immensely.
I started digging deep regarding Artificial Intelligence, Brain-inspired Computational Architectures, Neuromorphic Computing, and their future aspects. I was not well aware of the brain from a biological perspective. Still, I was dazzled by how complex and enormously efficient human brains are and how recent research took inspiration from brains to build a radically different computational architecture that is much faster and more efficient than any traditional design we have ever seen. I started doing extensive surveys and reading literature and realized that this field would change the fundamental ways in which we do computations and had a tremendous scope in the future. Hence I decided to carry out my research in this area rather than pursuing a job. For this, I carefully researched the best institutes/research labs in India where the research related to the brain-inspired computational fields is carried out. And to my surprise, there were only a few, yet most of them were radically different from the work they displayed on the website. Digging deeper and giving interviews in almost all older IITs and IISc while simultaneously talking to the lab supervisors, students working in the lab, I finally found a research lab in IISc which met all the criteria I had in mind. Since doctoral research is a long-term commitment, I had some predetermined criteria about the area in which I wanted to do research, the kind of freedom I wanted in the research, and the reach/ collaboration I could have with different industries and other reputed institutes. I think what really brought me to research is my willingness to keep on learning new things and decision not to be stuck in the loop where people do the same work repeatedly and forget about their growth just because they become comfortable in the area they are working on. I always wanted to contribute to society, and I realized this was also a way to contribute.
How did you plan the steps to get into the career you wanted? Or how did you make a transition to a new career? Tell us about your career path
I think the career path chosen by many individuals is either motivated by one’s interest in the subject matter or is driven by the monetary benefits. The prior one is a good choice in the long term while the latter one gives short-term satisfaction. In addition to that, I believe several other factors like family condition, family education, social environment, students’ willingness to learn, and their thought process play a key role in deciding the career path. For me, it was my initial experiences, from my 1st job in an MNC, my thought-process, my willingness to learn new things, and the complete support from my family that helped me decide on an unexplored yet auspicious career path. To put it into a sequence of events, I remember after completing my bachelor’s degree in Electronics and Communication, I had a number of on and off-campus placement offers in hand. Since I was also preparing for competitive examinations at that time (like most of us do as part of the rat-race), I decided to join an MNC with an expectation that I could get some time to prepare. But to my luck, there was none.That is the bane of a high performer.
During that time, automation software was booming, and everyone was willing to automate one or another process to cut down costs and efforts, but none was willing to invest extra money in it. During my work in MNC for the client, I took up the initiative to set up a complete automation framework. It was a year-long effort, and therefore, I didn’t have any time for exam preparation. The results of the projects were impressive. After the project was completed, I redid the same work for another client and then another. That’s when I realized that I was not making appropriate use of my knowledge. My work domain and area of interest were not matching up. So there was no point in continuing further as there was nothing new to learn. In the company, I excelled as a trainee, as a team member, and as a risk-taker as well. In a span of a year, I was responsible for automating end-to-end testing processes in multiple client accounts. I published two white papers in the same year regarding the architectural setup. That technology was still in the nascent stage, and I started getting offers from other companies as well. But, there was always a feeling that I could achieve much more than what I have achieved when I enhance the knowledge I possess in my field of interest and combine it with my full effort and dedication to achieve it. There is a saying that “There is a limit to what one can do if people themselves are limited”.
Along with my job, I decided to put in my sincere efforts in preparing for competitive exams and give it one more try. This helped me a lot in clearing my basics. I cleared UGS-NET-JRF, ESE prelims, and GATE multiple times. So, scholarship was not an issue for me. Then I left my job and joined the master’s program in VLSI at IIT. During the final year of my masters’ project under an experienced supervisor, I developed my interest in Brain-Inspired Computing and the related field of Neuromorphic Computing. I was dazzled by the human brain and how recent technological advancements took inspirations from human brains to build a radically different computational architecture that is faster and more efficient than any traditional/existing design. This interest led me to carry out research in this area rather than pursuing a job.
How did you get your first break?
I had multiple off-campus and on-campus offers before I graduated and after my master’s as well.
Ultimately I decided to choose a research-oriented path rather than a money-driven path. While researching, one develops a broad perspective on a wide variety of work and an understanding of oneself.
I believe a few years of experience in a company is required to understand the working environment and ethics.
What were the challenges you faced in your career? How did you address them?
I firmly believe that the problem one doesn’t want to address or keeps on postponing becomes a challenge sooner or later. It’s always wise to solve the problem instantly or make a sincere effort to do so. When I started my first job in 2014, I slowly started being comfortable with the typical MNC routine until I realized that I could never grow until my education is itself limited. Deciding between the right path (the path I loved to follow) and money was a difficult decision at that time. And I am grateful I made the right choice. I decided to give up my job to pursue a career in research. In all those moments, the support and encouragement from my family were of immense help. However, once I made up my mind, a tremendous effort was required to achieve it. Preparing for competitive exams while simultaneously working/delivering to your full potential in the company was a challenging one.
It might sound a bit odd, but It was also a challenge to keep my younger brother motivated while preparing for his competitive examinations. I felt it was/is my responsibility to guide him in the right path as my parents guided me. It was a year-long educational and motivational guidance that was required which finally paid off.
Where do you work now? Tell us about your research
Currently, I am pursuing research at Indian Institute of Science (IISc), Bangalore. The research projects I have been working on require a multi-disciplinary approach from a wide range of disciplines: Artificial Intelligence, Machine Learning, Analog, and mixed-signal circuit designs, device physics, and a bit of neurobiology. At one point, all these fields merge together with an end goal of designing highly efficient computational architectures that are biologically inspired by how human brains work . This research belongs to a class called Brain-Inspired Artificial Intelligence but not limited to Neuromorphic Computing. The definition of these fields is getting vague day by day due to its interdisciplinary nature. Firstly, to understand the research, it is important to define the objective and the problem one is working on. The computational architectures that I build are inspired by human brains. Why human brains? The human brain is far more advanced, efficient, and possesses more raw computational power than the most impressive supercomputers ever built. Even the fastest supercomputer in the world doesn’t come close to the processing speed of the human brain. As an approximation, the human brain operates at 1exaFLOP, which is equivalent to a billion billion calculations per second. The same calculations and processes that could take a supercomputer a few millions steps can be achieved by a few hundred neurons (basic computational unit) in the brain, requiring far less energy(~ 20 W) while performing at a far greater efficiency. In a vague comparison, the amount of energy required to power computations by the world’s fastest supercomputer would be enough to power a building or a complete cricket stadium; on the other hand, the human brain achieves the same processing speeds from the same energy that is required to charge a dim light bulb. Not to mention, supercomputers can be the size of a football field, but will still fail to be even comparable to this biological miracle. My work lies in bridging the gap between the computational efficiency of the brain and machine. I work on designing architectures that are far different from any traditional design and are inspired in a crude way by how the human brain works. Those architectures have for more computational capability that any similar architectures present today.
The end goal is to create a “state of the art” brain-inspired processor for artificial intelligence and machine learning tasks, whose performance parameters can beat today’s “state of the art” chips. Since the expectations are high and the field is relatively new, an immense effort and overall collaboration with different researchers in different fields from different universities is very much required.
To make my work more efficient and focused, I always try to set unrealistic goals (in terms of a part of work to be done or a particular research to be carried out) for me so that even if I fail, I am always much closer to success. Since this field is niche and new, I never get bored. There is always something new to learn and most of the time it’s the failure that teaches me the lesson. As a researcher, I have learnt to embrace failure and learn from it so as not to repeat it again while simultaneously using the lessons learnt in the next problem. Research itself means a problem which is partially defined or undefined. If we know everything from the beginning then it can never be called research.
How does your work benefit society?
Taking inspiration from the biological miracle and trying to bridge the gap between biology and machines is what I do. Carrying out cutting-edge research in Artificial Intelligence inspired by the brain and building architectures with an end product in mind that has the potential to revolutionize the present industry, is what I, as a researcher, intend to do. Nature teaches us a lot in one way or another. One can always take inspiration from nature and make an effort to solve one or many existing problems in today’s world.
As per the path I chose, and I mentioned above already, I was dazzled by how the human brain works and how I can utilize my existing knowledge in a specific field by creating something that could significantly impact society in one way or another.
Today is the era of digitization. We are generating so much data in a day that we have never generated in the last 50 years. Whether it is Facebook, Google, WhatsApp, Paytm, and anything digital, we need to make sense of all this data. The present technology is not enough to give us such high computation power to make sense of such data quickly. We wish we have human brains with super-fast computational speed to make sense of all this big data. My work focuses on taking inspiration from the brain and building Integrated circuits that can do cognitive tasks like the human brain. This will help in bridging the gap between humans and machines.
Tell us an example of a specific memorable work you did that is very close to you!
Every work gives us a learning opportunity. One of the things I remember is the project I had started a few years ago without any knowledge on what path to follow, or what could be the end result, or will it even succeed, and was it even worth putting effort into. But against all odds, I completed it. In the process, I remember designing multiple solutions and failing each time continuously for months. There were times when I gave up, but then there was always a belief and an unknown encouragement telling me “If I could just pull this off.” After multiple failures, I changed my approach to tackle the problem. Instead of getting things right, I started eliminating all the possibilities that could have resulted in the failure with the hope that one day whatever was left will make this project work. And to my surprise, after a few months of rectifying failures, it worked.
Your advice to students based on your experience?
The first piece of advice to all the students is that you should always plan your future well ahead, think at least 5 years ahead of your time, and remember there are no shortcuts in life; there has never been one. It is only when you start learning from fundamentals that you get to see the bigger picture, with immense possibility of innovations in them. One should always focus on learning and growth in this rapidly changing era, rather than short-term monetary benefits, to pave the path for a better future.
Secondly, one should always learn to prioritize their work/tasks. If you have ‘n’ tasks in your hand, prioritize them and then solve them from high priority to least priority.
Lastly, always learn to share and contribute. One can learn a lot more by sharing than just keeping the knowledge to oneself.
Since education is an ever-changing subject and the borders segregating disciplines are becoming blurred day by day, the technology that is dominating today may not even exist tomorrow. The only way to sustain is to keep on learning throughout life. Applying the same principle to myself, my hope is to continue doing cutting-edge research in the ever-learning and growing field of artificial intelligence, machine learning, brain-inspired computations, and related fields. Simultaneously, the resulting end product should also benefit humanity and society.