“After my B.Tech in Electrical Engineering, I thought of going for an MSc program at IISc, Bangalore. But instead of applying to the Electrical Engineering department, the subject that I found interesting and applied for was the BioSystems Science and Engineering Department, which I didn’t realise till I received my interview call letter, had only a PhD programme and not a Master’s programme. So my entry into PhD was somewhat serendipitous.
I worked on systems neuroscience with a focus on modelling our visual system. My project aimed at studying the changes in the brain associated with reading, especially while encountering jumbled words, alpha-numeric expressions, etc, using behavioural and brain imaging (fMRI) experiments .
During my B.Tech, I was fascinated about neural networks/artificial intelligence and joined this PhD programme with a goal to reverse engineer the brain and build intelligent machines. I thought I would do more machine learning but the reality differed quite a lot from that expectation. Transitioning from engineering to research in systems biology was not easy. I had to do more experimental work, which I still do not enjoy and try to avoid it whenever possible. After I got in the program, it took me two semesters in the beginning to understand and correctly interpret terminologies associated with biological research. I faced a major setback at the end of second year when all my experiments failed and I had to entirely switch my research topic. Fortunately, my guides were open to let me pursue my interests. However, the topic I chose was not their core expertise. They specialised in vision and signal processing but my chosen project involved another aspect of our cognitive ability i.e. language processing Though they were open to exploring this research area, we had to put a lot of effort when we were starting out on literature survey, analysis and also in getting my first paper published.
The later part of the second year of my PhD was my lowest point. My experimental findings were not conclusive. In one incident, I put a person under an MRI scanner to record his brain signals but I ignored a few crucial steps due to which I had to abort my experiment. My supervisor was unhappy with me due to my negligence. Also, since we needed human volunteers to do many behavioural experiments, it took a toll on my personal time and social life because most of them were available outside our working hours either on Sundays, other holidays or even late nights. So the experimental work was quite exhausting. The only part of the research I thoroughly enjoyed was analysing the data, building predictive models and proposing new theories.
After considerable hard work and a patient wait, my first paper finally got published in 2019, I submitted my thesis in August 2019 and defended it in January 2020. Post my defense, I tried my hands at entrepreneurship for a while. I was selected as founder-in-residence at Entrepreneur First, which is a boot camp of sorts where people with different expertise come together to conceptualize ideas, form teams and start a company. At the end of three months we had to pitch our idea to an expert panel who would shortlist a billion-dollar idea and invest on it. But with the announcement of lockdown in March, we couldn’t satisfy all criteria needed to get funding; so that endeavour did not materialise.
All my research work and experiments so far have been small steps and learning processes to realise my goal of creating an efficient artificial intelligence (AI) system that would one day match visual abilities such as reading handwritten characters, distorted strings such as CAPTCHAs etc. Currently I am pursuing a Postdoctoral research in Paris, where I plan to conduct research that will bring me closer towards materializing my dreams.
Maybe my research doesn’t have direct industrial applications but the understanding of brain functioning will help develop robust AI systems which in turn can have diverse applications such as interpreting the actions of amputees and helping them artificially to coordinate their limb activities.
India is still lacking behind in adopting latest technologies, maybe due to availability of cheap labor. Whether AI revolution will replace human resources is a debatable issue. On a positive side, we can consider it as an advancement in technology that increases convenience- much like how we look at mobile phones as a tool of convenience rather than being a replacement to the job of a postman. Use of technology is relative and I don’t think it fully threatens a human resource intensive society.”
-Aakash Agarwal, PhD in Biosystems Science and Engineering, Indian Institute of Science, Bangalore