Events Calendar

EE599: Causal Learning course: Casual Bandits

Speaker: Ali Tajer, Rensselaer Polytechnic Institute

Talk Title: Causal Bandits

Abstract: In this talk, we provide an overview of the causal bandit problems. The purpose of causal bandit settings is to formalize theoretically-principled frameworks for the experimental design when the experiments involve an array of parameters that causally affect one another. The key objective of causal bandits is to leverage causal relationships to design effective experiments judiciously. Designing causal bandit algorithms critically hinges on the extent of information available about the (i) causal structure and (ii) the interventional distributions. Based on the availability of information on each of these two dimensions, there are, broadly, four possible model combinations. The existing literature, for the most part, focuses on settings in which the interventional distributions are known (with or without knowing the causal structure). First, we provide an overview of the existing literature on the existing literature. Secondly, motivated by the fact that acquiring the interventional distributions is often infeasible, we address the following question: is it possible to achieve the optimal regret scaling rates without knowing the interventional distributions? We address this question affirmatively in the case of linear structural equation models when the causal structure is known. We discuss the design and performance of algorithms for the frequentist and Bayesian settings.

Biography: Ali Tajer received the B.Sc. and M.Sc. degrees in Electrical Engineering from Sharif University of Technology in 2002 and 2004, respectively. During 2007-2010 he was with Columbia University, where he received an M.A degree in Statistics and a Ph.D. degree in Electrical Engineering, and during 2010-2012 he was with Princeton University as a Postdoctoral Research Associate. He is currently an Associate Professor of Electrical, Computer, and Systems Engineering at Rensselaer Polytechnic Institute. His research interests include mathematical statistics, statistical signal processing, and network information theory, with applications in wireless communications and power grids. His recent publications include an edited book entitled Advanced Data Analytics for Power Systems (Cambridge University Press, 2021). He received an NSF CAREER award in 2016 and AFRL Faculty Fellowship in 2019. He is currently serving as an Associate Editor for the IEEE Transaction on Information Theory and an Associate Editor for the IEEE Transactions on Signal Processing. In the past, he has served as an Editor for the IEEE Transactions on Communications, a Guest Editor for the IEEE Signal Processing Magazine, an Editor for the IEEE Transactions on Smart Grid, an Editor for the IET Transactions on Smart Grid, and as a Guest Editor-in-Chief for the IEEE Transactions on Smart Grid -“ Special Issue on Theory of Complex Systems with Applications to Smart Grid Operations.

Host: Urbashi Mitra; Password for link: 114454

More Info: https://usc.zoom.us/j/94255391488?pwd=cGoyOVoxWnc3K1RTeVcvYjlWOEJPQT09

Location: Virtual

Audiences: Everyone Is Invited

Contact: Susan Wiedem

Event Link: https://usc.zoom.us/j/94255391488?pwd=cGoyOVoxWnc3K1RTeVcvYjlWOEJPQT09