Weinan E
Academician, Chinese Academy of Sciences
Michael I. Jordan
Member, National Academy of Sciences Member, National Academy of Engineering Member, American Academy of Arts and Sciences
Michael P. Wellman
Professor, University of Michigan
Competition and Cooperation in LLM Bargaining: An Empirical Meta-Game Analysis
Abstract
Machine learning algorithms and LLM agents with interactive capabilities can be viewed as meta-strategies, mapping environment descriptions and training experience to strategies for interacting in particular situations. I present a framework that evaluates such techniques through induction of an empirical game model from simulation data. The approach is illustrated in the setting of two-player bargaining, applied to multiagent reinforcement learning algorithms as well as LLM agents. The analysis helps illuminate the potential tension between competition and cooperation in a canonical multiagent interaction setting.
Biography
Michael P. Wellman is Lynn A. Conway Collegiate Professor of Computer Science & Engineering at the University of Michigan. He received a PhD from the Massachusetts Institute of Technology in 1988 for his work in qualitative probabilistic reasoning and decision-theoretic planning. From 1988 to 1992, Wellman conducted research in these areas at the USAF’s Wright Laboratory.
For the past 30+ years, his research has focused on computational market mechanisms and game-theoretic reasoning methods, with applications in electronic commerce, finance, and cyber-security. As Chief Market Technologist for TradingDynamics, Inc., he designed configurable auction technology for dynamic business-to-business commerce. Wellman previously served as Chair of the ACM Special Interest Group on Electronic Commerce (SIGecom), and as Executive Editor of the Journal of Artificial Intelligence Research. He is a Fellow of the Association for the Advancement of Artificial Intelligence and the Association for Computing Machinery.