Download as iCal file
By Omer Ben Porat
Location Bloomfield 152, Faculty of Industrial Engineering and Management
Advisor(s): Moshe Tennenholtz
Academic Program: Please choose
Wednesday 13 November 2019, 11:30 - 12:30
Recommendation systems often face exploration-exploitation tradeoffs: the system can only learn about the desirability of new options by recommending them to some user. Such systems can thus be modeled as multi-armed bandit settings; however, users are self-interested and cannot be enforced to use the system. We ask whether exploration can nevertheless be performed in a way that scrupulously respects agents' interests---i.e., by a system that acts as a emph{fiduciary}.
More formally, we introduce a  model in which a system faces an exploration-exploitation tradeoff under the constraint that it can never recommend any action that it knows yields lower reward in expectation than an agent would achieve if it acted alone. Our main contribution is a positive result: an asymptotically optimal, incentive compatible, and emph{ex-ante} individually rational recommendation algorithm. 
The talk is based on joint work with Gal Bahar, Kevin Leyton-Brown, and Moshe Tennenholtz, which is available here.
Short Bio:
Omer Ben-Porat is a fifth-year PhD student at the Industrial Engineering and Management faculty of the Technion, working with Professor Moshe Tennenholtz. Omer's research interests lie in the overgrowing intersection of Machine Learning and Game Theory. In his PhD, he focuses on strategic aspects of Data Science, and have recently co-organized a workshop on that topic at the ACM FCRC 2019. Omer is a recipient of a J.P. Morgan PhD Fellowship 2019.