Predictive power of game theory for intelligent and autonomous systems

Departments: Electrical Engineering
Time: Tuesday, February 23, 2021 - 4:00pm - 5:00pm
Type: Seminar Series
Presenter: Muhammed Omer Sayin; Laboratory for Information and Decision Systems (LIDS), MIT
Room/Office:
Location:

Department of Electrical Engineering Seminar

"Predictive power of game theory for intelligent and autonomous systems"

Muhammed Omer Sayin
Laboratory for Information and Decision Systems (LIDS), MIT

Tuesday, February 23, 2021 4:00 PM

Abstract: ​From self-driving vehicles to Internet-of-Things devices, intelligent and autonomous systems are becoming integral to everyday life. These systems can now achieve cognitive abilities we once knew as innately human such as perception and rationalization. The prevalence of these intelligent systems in our society brings new research opportunities to enhance and re-envision their interactions with each other and humans so that we can harness their potential altogether. This talk will focus on developing principled understanding and systematic guarantees for their behaviors in dynamic and complex environments. We will present a new solution concept built on the predictive power of game theory and behavioral dynamics. More explicitly, we will show that their behaviors converge to equilibrium in a particular class of dynamic games called zero-sum stochastic games if they follow a simple natural learning dynamic known as fictitious play. This was a long-standing problem in the theory of learning in games going back to George W. Brown and Lloyd S. Shapley's seminal papers on the fictitious play and stochastic games in the early 1950s. Furthermore, this solution concept draws a framework bridging single-agent and multi-agent learning to study the near future's emerging technologies and applications. Finally, we will present our research vision and future research agenda directed towards developing human-like social intelligence.

Bio:​ Muhammed Omer Sayin is a Postdoctoral Associate at the Laboratory for Information and Decision Systems (LIDS), Massachusetts Institute of Technology (MIT), and advised by Asuman Ozdaglar. He got his Ph.D. from the University of Illinois at Urbana-Champaign (UIUC) in 2019. His advisor was Tamer Başar. During his Ph.D., he had two research internships in Toyota InfoTech Labs, Mountain View, CA. He got his M.S. and B.S. from Bilkent University, Turkey, respectively, in 2015 and 2013. The overarching theme of his research is to develop solution concepts that can forecast and control the behavior of multiple intelligent and autonomous systems via the predictive power of game theory.​

Hosted by:​ Prof. Steve Morse

Contact the department admins for Zoom information.