Think Globally, Act Locally: Distributed Learning Algorithms for Cooperative and Interactive Agents

Ilai Bistritz, PhD Candidate in the Electrical Engineering department at Stanford University

​Tel Aviv University​, 29 March 2022, 14:00 PM, Room 206& via zoom

Abstract:
Machine learning aims to relegate many decision-making processes to agents (machines). In applications like smart cities, autonomous vehicles, the Internet of Things, and robotics, the agents interact in a networked environment, where the decisions of one agent affect the other agents and their learning. As game theory predicts, this interaction can lead to inefficient outcomes for all agents involved. However, machines follow programmatic objectives and protocols that, unlike humans, are not limited by selfish interests but by information and resources. This modern paradigm calls for new tools to design efficient multi-agent protocols. In this talk, we will discuss such recently developed tools and future challenges in unleashing the full potential of this paradigm.

Bio:
Ilai Bistritz is a PhD Candidate in the Electrical Engineering department at Stanford University, advised by Prof. Nicholas Bambos. He received the B.Sc. (magna cum laude) and M.S. (summa cum laude) degrees in electrical engineering from Tel-Aviv University in 2012 and 2016, respectively. His main research interests are game theory, distributed control, and multi-agent learning.