Actor-Critic Fictitious Play in Simultaneous ...
Type de document :
Communication dans un congrès avec actes
Titre :
Actor-Critic Fictitious Play in Simultaneous Move Multistage Games
Auteur(s) :
Pérolat, Julien [Auteur]
Sequential Learning [SEQUEL]
Université de Lille, Sciences et Technologies
Piot, Bilal [Auteur]
IMS : Information, Multimodalité & Signal
Pietquin, Olivier [Auteur]
IMS : Information, Multimodalité & Signal
Sequential Learning [SEQUEL]
Université de Lille, Sciences et Technologies
Piot, Bilal [Auteur]
IMS : Information, Multimodalité & Signal
Pietquin, Olivier [Auteur]
IMS : Information, Multimodalité & Signal
Titre de la manifestation scientifique :
AISTATS 2018 - 21st International Conference on Artificial Intelligence and Statistics
Ville :
Playa Blanca, Lanzarote, Canary Islands
Pays :
Espagne
Date de début de la manifestation scientifique :
2018-04-09
Discipline(s) HAL :
Informatique [cs]
Informatique [cs]/Intelligence artificielle [cs.AI]
Informatique [cs]/Intelligence artificielle [cs.AI]
Résumé en anglais : [en]
Fictitious play is a game theoretic iterative procedure meant to learn an equilibrium in normal form games. However, this algorithm requires that each player has full knowledge of other players' strategies. Using an ...
Lire la suite >Fictitious play is a game theoretic iterative procedure meant to learn an equilibrium in normal form games. However, this algorithm requires that each player has full knowledge of other players' strategies. Using an architecture inspired by actor-critic algorithms, we build a stochastic approximation of the fictitious play process. This procedure is on-line, decentralized (an agent has no information of others' strategies and rewards) and applies to multistage games (a generalization of normal form games). In addition, we prove convergence of our method towards a Nash equilibrium in both the cases of zero-sum two-player multistage games and cooperative multistage games. We also provide empirical evidence of the soundness of our approach on the game of Alesia with and without function approximation.Lire moins >
Lire la suite >Fictitious play is a game theoretic iterative procedure meant to learn an equilibrium in normal form games. However, this algorithm requires that each player has full knowledge of other players' strategies. Using an architecture inspired by actor-critic algorithms, we build a stochastic approximation of the fictitious play process. This procedure is on-line, decentralized (an agent has no information of others' strategies and rewards) and applies to multistage games (a generalization of normal form games). In addition, we prove convergence of our method towards a Nash equilibrium in both the cases of zero-sum two-player multistage games and cooperative multistage games. We also provide empirical evidence of the soundness of our approach on the game of Alesia with and without function approximation.Lire moins >
Langue :
Anglais
Comité de lecture :
Oui
Audience :
Internationale
Vulgarisation :
Non
Collections :
Source :
Fichiers
- https://hal.inria.fr/hal-01724227/document
- Accès libre
- Accéder au document
- https://hal.inria.fr/hal-01724227/document
- Accès libre
- Accéder au document
- https://hal.inria.fr/hal-01724227/file/AISTATS2018%20%2813%29.pdf
- Accès libre
- Accéder au document
- document
- Accès libre
- Accéder au document
- AISTATS2018%20%2813%29.pdf
- Accès libre
- Accéder au document
- document
- Accès libre
- Accéder au document
- AISTATS2018%20%2813%29.pdf
- Accès libre
- Accéder au document