Title: Mechanism Design for Stochastic Multi-armed Bandit Problems
Speaker:Dr Shweta Jain, IIT Bhubaneshwar
Host Faculty: Dr. Maria Francis
Room No: A-119, academic block A
Time: 2:30 - 3:30
The multi-armed bandit (MAB) problem is widely studied in the machine learning literature, in the context of online learning. My recent work has studied the MAB problem when the arms are controlled by strategic agents. Dealing with strategic agents warrants the use of principled tools such as game theory and mechanism design in conjunction with online learning, and leads to non-trivial technical challenges. My research offers novel mechanisms for stochastic MAB problems in the various representative, generic settings inspired by applications such as smart grids, crowdsourcing, Internet advertising, and procurement auction. In this talk, I will discuss in detail about one particular problem of designing a low regret, combinatorial, randomized mechanism for the MAB problem, where, we consider a problem of selecting a subset of agents to solve a general combinatorial optimization problem.
Dr Shweta Jain is a Ph.D. from the Department of Computer Science and Automation (CSA), Indian Institute of Science. She earlier completed her M.E. from CSA and had worked for a year at Qualcomm. During her Ph.D., she explored mechanism design problems in the context of online learning, using the multi-armed bandit abstraction. She is a recipient of the Microsoft Doctoral Fellowship during her Ph.D. Currently, she is a faculty at IIT Bhubaneshwar. Her research interests include Game Theory, Mechanism Design, Machine Learning, Applications to Auctions, Crowdsourcing, Internet Advertising, and Artificial Intelligence.
Dates: Thursday, August 23, 2018 - 12:00 to 13:00