Fair Machine Learning via Information Theory (1/3)
Presenter(s)
Presenter Profile Picture
KAIST

2021 Croucher Summer Course in Information Theory, The Chinese University of Hong Kong

Lecture

Date

Abstract

We explore an interesting topic in the field of Artificial Intelligence (AI) via the lens of information theory: fair machine learning. Fairness is one of the crucial aspects required for enabling trustworthy AI. In this series of lectures, we study how tools of information theory serve to develop fair classifiers that aim to achieve the irrelevancy of a prediction to sensitive attributes such as race, sex, age and religion. We also investigate an intimate connection to one prominent unsupervised learning framework: generative adversarial networks.

Biography
Changho Suh is an Associate Professor and a Vice-Chair of Electrical Engineering at KAIST. He received the B.S. and M.S. degrees in Electrical Engineering from KAIST in 2000 and 2002 respectively, and the Ph.D. degree in EECS from UC Berkeley in 2011. From 2011 to 2012, he was a postdoctoral associate in MIT. From 2002 to 2006, he had been with Samsung. Prof. Suh is a recipient of numerous awards: the AFOSR Grant, the Google Education Grant, the Young IT Engineer Award, the IEIE Haedong Young Engineer Award, the IEEE Communications Society Stephen O. Rice Prize, the David J. Sakrison Memorial Prize (the best dissertation award in UC Berkeley EECS), the IEEE ISIT Best Student Paper Award, and the Department Teaching Awards. Dr. Suh is an IEEE Information Theory Society Distinguished Lecturer, the General Co-Chair of East Asian School of Information Theory, and a Member of Young Korean Academy of Science and Technology. He is also the Editor for the IEEE Information Theory Newsletter, an Associate Editor for Statistical Learning for the IEEE Transactions on Information Theory, and a Senior Program Committee member of IJCAI 2019–2021.