Learning Nash Equilibrium for General-Sum Markov Games from Batch Data

Julien Pérolat 1, 2 Florian Strub 2 Bilal Piot 3 Olivier Pietquin 1, 4, 3, 2
1 Université de Lille, Sciences et Technologies
2 SEQUEL - Sequential Learning
Inria Lille - Nord Europe, CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille (CRIStAL) - UMR 9189
3 DeepMind
4 IUF - Institut Universitaire de France
Abstract : This paper addresses the problem of learning a Nash equilibrium in γ-discounted mul-tiplayer general-sum Markov Games (MGs) in a batch setting. As the number of players increases in MG, the agents may either collaborate or team apart to increase their final rewards. One solution to address this problem is to look for a Nash equilibrium. Although , several techniques were found for the subcase of two-player zero-sum MGs, those techniques fail to find a Nash equilibrium in general-sum Markov Games. In this paper, we introduce a new definition of-Nash equilibrium in MGs which grasps the strategy's quality for multiplayer games. We prove that minimizing the norm of two Bellman-like residuals implies to learn such an-Nash equilibrium. Then, we show that minimizing an empirical estimate of the L p norm of these Bellman-like residuals allows learning for general-sum games within the batch setting. Finally, we introduce a neural network architecture that successfully learns a Nash equilibrium in generic multiplayer general-sum turn-based MGs.
Document type :
Conference papers
Complete list of metadatas

Cited literature [26 references]  Display  Hide  Download
https://hal.inria.fr/hal-01648489
Contributor : Julien Pérolat <>
Submitted on : Sunday, November 26, 2017 - 1:37:11 PM
Last modification on : Thursday, April 4, 2019 - 10:18:05 AM

File

bellman-residual-aistats2016(5...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01648489, version 1

Collections

CNRS | INRIA | CRISTAL-SEQUEL | CRISTAL | UNIV-LILLE

Citation

Julien Pérolat, Florian Strub, Bilal Piot, Olivier Pietquin. Learning Nash Equilibrium for General-Sum Markov Games from Batch Data. AISTATS 2017 - The 20th International Conference on Artificial Intelligence and Statistics, Apr 2017, Fort Lauderdale, United States. pp.1-14. ⟨hal-01648489⟩

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Record views

196

Files downloads

179

 

 

Contact

archive-ouverte@inria.fr