Master Equation for Discrete-Time Stackelberg Mean Field Games
From Google TechTalks · 43:02
Show by Google TechTalks
A Google TechTalk, presented by Deepanshu Vasal, 2022/09/13A Google Algorithms Seminar - ABSTRACT: Mean field games are a very popular model of dynamic interaction of large scale selfish agents. In this talk, I will consider a general model of discrete-time Stackelberg mean field games with a leader and an infinite number of followers. The leader and the followers each observe types privately that evolve as conditionally independent controlled Markov processes. The leader commits to a dynamic policy and the followers best respond to that policy and each other. Knowing that the followers would play a mean field game based on her policy, the leader chooses a policy that maximizes her reward. We refer to the resulting outcome as a Stackelberg mean field equilibrium (SMFE).Uploaded 20-09-22In this talk, I will provide a master equation of this game that allows one to compute all SMFE backward recursively. Based on this framework, I will consider two numerical examples, one on vaccinations in an epidemic and other on technology adoption. This presents a new idea and a framework in solving such games and has vast potential in designing strategies of the government and firms in control of societal level problems including information design, pricing, vaccination costs, blockchains and much more. About the Speaker:Deepanshu Vasal is a research scientist in the Department of Electrical and Computer Engineering (ECE) at Northwestern University. He received his PhD from University of Michigan, Ann Arbor in EE:Systems in 2016, and was a postdoc at UT Austin. His current research interests are game theory, multi-agent decision making, and feedback communication. Before that, he received his B.Tech degree in electronics and communication engineering from IIT Guwahati in 2009.
More episodes of Google TechTalks
More episodes
- DSD: A Hybrid Analysis Tool for Bug Finding
- Chasing the Constants and its Implications in Differential Privacy
- Making Great Decisions
- Streaming Private Continual Counting via Binning
- Distributed Testing with SmartFrog
- Differentially Private Synthetic Data without Training
- A Dynamic Network Layer for Advanced Cloud Computing
- Data Representation/Laplace Operator
- Persistent Pre-Training Poisoning of LLMs
- Threat Models for Memorization: Privacy, Copyright, and Everything In-Between
PodSearch is a free, independent podcast search engine. Read about PodSearch, the privacy policy or contact us.