Passer à la navigation principale Passer à la recherche Passer au contenu principal

Towards Minimax Optimality of Model-based Robust Reinforcement Learning

  • Ecole polytechnique
  • HeKA
  • Centre de Recherche des Cordeliers
  • Cohere

Résultats de recherche: Contribution à un journalArticle de conférenceRevue par des pairs

Résumé

We study the sample complexity of obtaining an ϵ-optimal policy in Robust discounted Markov Decision Processes (RMDPs), given only access to a generative model of the nominal kernel. This problem is widely studied in the non-robust case, and it is known that any planning approach applied to an empirical MDP estimated with Õ(H3 |ϵS2||A|) samples provides an ϵ-optimal policy, which is minimax optimal. Results in the robust case are much more scarce. For sa- (resp s-) rectangular uncertainty sets, until recently the best-known sample complexity was Õ(H4 |Sϵ2|2 |A|) (resp. Õ(H4 |Sϵ|22 |A|2)), for specific algorithms and when the uncertainty set is based on the total variation (TV), the KL or the Chi-square divergences. In this paper, we consider uncertainty sets defined with an Lp-ball (recovering the TV case), and study the sample complexity of any planning algorithm (with high accuracy guarantee on the solution) applied to an empirical RMDP estimated using the generative model. In the general case, we prove a sample complexity of Õ(H4 |ϵS2||A|) for both the sa- and s-rectangular cases (improvements of |S| and |S||A| respectively). When the size of the uncertainty is small enough, we improve the sample complexity to Õ(H3 |ϵS2||A|), recovering the lower-bound for the non-robust case for the first time and a robust lower-bound. Finally, we also introduce simple and efficient algorithms for solving the studied Lp robust MDPs.

langue originaleAnglais
Pages (de - à)820-855
Nombre de pages36
journalProceedings of Machine Learning Research
Volume244
étatPublié - 1 janv. 2024
Evénement40th Conference on Uncertainty in Artificial Intelligence, UAI 2024 - Barcelona, Espagne
Durée: 15 juil. 202419 juil. 2024

Empreinte digitale

Examiner les sujets de recherche de « Towards Minimax Optimality of Model-based Robust Reinforcement Learning ». Ensemble, ils forment une empreinte digitale unique.

Contient cette citation