Abstract

Representations are functions of data that enable us to distill the information present within. Ideally, they should be as informative as the raw data, unaffected by nuisance factors in unseen data, as simple as possible, and easy to work with. Such ideal representations can then be stored in memory in lieu of raw data and utilized for downstream tasks. Motivated by these desiderata, a significant amount of effort has been devoted to designing effective representations over the past century, starting from principal component analysis and culminating in the current day paradigm of data-driven representation learning.

Bio

Vishnu Naresh Boddeti is currently an Assistant Professor of Computer Science and Engineering at Michigan State University. He received a Bachelor's degree in Electrical Engineering in 2003 from the Indian Institute of Technology, Madras. He received a Ph.D. degree in Electrical and Computer Engineering from Carnegie Mellon University in 2013. His research interests are in Machine Learning, Computer Vision and Signal Processing. He received the best paper award at the BTAS conference in 2013, the best student paper award at the ACCV conference in 2018 and the best paper award in evolutionary machine learning at the GECCO conference in 2019.

Host
Professor Xiaoqian (Joy) Wang, joywang@purdue.edu