Learning Game Representations from Data Using Rationality Constraints
Xi Gao, Avi Pfeffer
While game theory is widely used to model strategic interactions, a natural question is where do the game representations come from? One answer is to learn the representa- tions from data. If one wants to learn both the payoffs and the players' strategies, a naive approach is to learn them both directly from the data. This approach ignores the fact the players might be playing reasonably good strategies, so there is a connection between the strategies and the data. The main con- tribution of this paper is to make this connec- tion while learning. We formulate the learn- ing problem as a weighted constraint satis- faction problem, including constraints both for the fit of the payoffs and strategies to the data and the fit of the strategies to the pay- os. We use quantal response equilibrium as our notion of rationality for quantifying the latter fit. Our results show that incorporat- ing rationality constraints can improve learn- ing when the amount of data is limited.
PDF Link: /papers/10/p185-gao.pdf
AUTHOR = "Xi Gao
and Avi Pfeffer",
TITLE = "Learning Game Representations from Data Using Rationality Constraints",
BOOKTITLE = "Proceedings of the Twenty-Sixth Conference Annual Conference on Uncertainty in Artificial Intelligence (UAI-10)",
PUBLISHER = "AUAI Press",
ADDRESS = "Corvallis, Oregon",
YEAR = "2010",
PAGES = "185--192"