Regret Bounds for Learning State Representations in Reinforcement Learning

Ortner, Ronald; Pirotta, Matteo; Fruit, Ronan; Lazaric, Alessandro; Maillard, Odalric Ambrym

Type de document :

Communication dans un congrès avec actes

Titre :

Regret Bounds for Learning State Representations in Reinforcement Learning

Auteur(s) :

Ortner, Ronald [Auteur]
Montanuniversität Leoben [MUL]
Pirotta, Matteo [Auteur]
Facebook AI Research [Paris] [FAIR]
Fruit, Ronan [Auteur]
Sequential Learning [SEQUEL]
Lazaric, Alessandro [Auteur]

Facebook AI Research [Paris] [FAIR]
Maillard, Odalric Ambrym [Auteur]

Sequential Learning [SEQUEL]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]

Titre de la manifestation scientifique :

Conference on Neural Information Processing Systems

Ville :

Vancouver

Pays :

Canada

Date de début de la manifestation scientifique :

2019-12

Titre de la revue :

Conference on Neural Information Processing Systems

Discipline(s) HAL :

Statistiques [stat]/Machine Learning [stat.ML]
Mathématiques [math]/Statistiques [math.ST]

Résumé en anglais : [en]

We consider the problem of online reinforcement learning when several state representations (mapping histories to a discrete state space) are available to the learning agent. At least one of these representations is assumed ...
Lire la suite >We consider the problem of online reinforcement learning when several state representations (mapping histories to a discrete state space) are available to the learning agent. At least one of these representations is assumed to induce a Markov decision process (MDP), and the performance of the agent is measured in terms of cumulative regret against the optimal policy giving the highest average reward in this MDP representation. We propose an algorithm (UCB-MS) with O(√ T) regret in any communicating MDP. The regret bound shows that UCB-MS automatically adapts to the Markov model and improves over the currently known best bound of order O(T 2/3).Lire moins >

Langue :

Anglais

Comité de lecture :

Oui

Audience :

Internationale

Vulgarisation :

Non

Projet ANR :

BANDITS MANCHOTS POUR SIGNAUX NON-STATIONNAIRES ET STRUCTURES

Collections :