AI Research Scholar’s Day 2021

This year, the Department of Artificial Intelligence celebrated its first Research Scholar's Day, which various professors from different departments attended. The first Ph.D. batch of the AI department, started in 2019, consisting of 3 students: Bhattaracharyulu V V R Dittakavi, Tejasri Nampally, and Suresh N, who presented their research as a part of the proceedings.

The event witnessed keynotes from Professor M. Vidyasagar, Distinguished Professor, IITH, and Prof. SP Arun, IISc. Professor Vidyasagar received his B.S., M.S., and Ph.D. degrees in electrical engineering from the University of Wisconsin in Madison in 1965, 1967, and 1969. He has received several awards in recognition of his research contributions, including Fellowship in The Royal Society, the world's oldest scientific academy in continuous existence, the IEEE Control Systems (Technical Field) Award, the Rufus Oldenburger Medal of ASME, the John R. Ragazzini Education Award from AACC, and others. Since March 2020, he is a SERB National Science Chair, one of four in India.

Prof. Vidyasagar’s talk covered his perspectives on Deep Learning and Reinforcement Learning. He firmly believes that advances in these fields backed by strong mathematical foundations will survive the test of time, and the rest would not. Prof. Vidyasagar began with a holistic overview of AI, covering the various sub-fields, the evolution of AI, and the overall goals. The remainder of the talk was mainly divided into two main sections: DL and RL.

Prof. SP Arun received his B.Tech in Electrical Engineering from IIT Bombay and MS and Ph.D. in Electrical Engineering from Johns Hopkins University. He completed his postdoctoral research at Carnegie Mellon University. He is currently an Associate Professor at the Centre of Neuroscience at IISC. His group works on how the brain transforms sensations into perceptions.

Prof. Arun spoke about improving machine vision using insights from biological vision. Initially, he compared machines and humans on hard tasks, bringing out systematic differences in error patterns. He then compared the representations in machines and humans, revealing that humans have systematic biases like symmetry. Lastly, he compared them based on perception and concluded that there are qualitative differences in perception. He hopes that by describing such differences, they can be incorporated into the machines to improve their performance.