Open AccessProceedings Article
Learning Simulation-Based Games from Data
Enrique Areyan Viqueira,Amy Greenwald,Cyrus Cousins,Eli Upfal +3 more
- 08 May 2019
- pp 1778-1780
10
TL;DR: Algorithms are designed that learn empirical games, which uniformly approximate the utilities of simulation-based games from finitely many samples, which learns all the equilibria of simulation -based games, as opposed to a single one.
read more
Abstract: We tackle a fundamental problem in empirical game-theoretic analysis (EGTA), that of learning equilibria of simulation-based games. Such games cannot be described in analytical form; instead, a black-box simulator can be queried to obtain noisy samples of utilities. Our approach to EGTA is in the spirit of probably approximately correct learning. We design algorithms that learn empirical games, which uniformly approximate the utilities of simulation-based games from finitely many samples. Our methodology learns all the equilibria of simulation-based games, as opposed to a single one.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
•Proceedings Article
Improved Algorithms for Learning Equilibria in Simulation-Based Games
Enrique Areyan Viqueira,Cyrus Cousins,Amy Greenwald +2 more
- 05 May 2020
TL;DR: It is shown that GS using the authors' variancesensitive bounds outperforms previous work, and that PSP can significantly outperform GS, which “outperform” means achieving the same guarantees with far fewer samples.
15
•Posted Content
Learning Probably Approximately Correct Maximin Strategies in Simulation-Based Games with Infinite Strategy Spaces.
TL;DR: This work designs two algorithms with theoretical guarantees to learn maximin strategies in two-player zero-sum games with infinite strategy spaces, and formally proveselta-PAC theoretical guarantees for these algorithms under some regularity assumptions.
14
•Proceedings Article
Sharp uniform convergence bounds through empirical centralization
Cyrus Cousins,Matteo Riondato +1 more
- 01 Jan 2020
TL;DR: The use of empirical centralization is introduced to derive novel practical, probabilistic, sample-dependent bounds to the Supremum Deviation of empirical means of functions in a family from their expectations, which greatly outperform non-centralized bounds and are extremely practical even at small sample sizes.
•Proceedings Article
Empirical Mechanism Design: Designing Mechanisms from Data.
Enrique Areyan Viqueira,Cyrus Cousins,Yasser Mohammad,Amy Greenwald +3 more
- 01 Jan 2019
TL;DR: This work introduces a methodology for the design of parametric mechanisms, which are multiagent systems inhabited by strategic agents, with knobs that can be adjusted to achieve specific goals, and incorporates the noise associated with modern concentration inequalities into the underlying Gaussian process.
Computational and Data Requirements for Learning Generic Properties of Simulation-Based Games
TL;DR: This work introduces a novel algorithm—progressive sampling with pruning (PSP)—for learning a uniform approximation and thus any well-behaved property of a game, which prunes strategy profiles once the corresponding players’ utilities are well-estimated, and analyzes its data and query complexities in terms of the a priori unknown utility variances.
1
References
Probability Inequalities for sums of Bounded Random Variables
TL;DR: In this article, upper bounds for the probability that the sum S of n independent random variables exceeds its mean ES by a positive number nt are derived for certain sums of dependent random variables such as U statistics.
A theory of the learnable
Leslie G. Valiant
- 05 Nov 1984
TL;DR: This paper regards learning as the phenomenon of knowledge acquisition in the absence of explicit programming, and gives a precise methodology for studying this phenomenon from a computational viewpoint.
A class of games possessing pure-strategy Nash equilibria
TL;DR: In this paper, a class of noncooperative games (of interest in certain applications) is described and each game in the class is shown to possess at least one Nash equilibrium in pure strategies.
2.4K
•Proceedings Article
Methods for empirical game-theoretic analysis
Michael P. Wellman
- 16 Jul 2006
TL;DR: An emerging empirical methodology bridges the gap between game theory and simulation for practical strategic reasoning in strategy and decision-making.
Stochastic search methods for nash equilibrium approximation in simulation-based games
Yevgeniy Vorobeychik,Michael P. Wellman +1 more
- 12 May 2008
TL;DR: A convergent algorithm based on a hierarchical application of simulated annealing for estimating Nash equilibria in simulation-based games with finite-dimensional strategy sets and considerable evidence that a method based on random search outperforms gradient descent in this setting is provided.
Related Papers (5)
Haris Aziz,Felix Brandt,Paul Harrenstein +2 more
- 10 May 2010
Nicolas Loizou
- 09 Dec 2016
Krishnendu Chatterjee,Nathanaël Fijalkow +1 more
- 01 Jun 2011