Only Relevant Information Matters: Filtering Out Noisy Samples to Boost RL - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

Only Relevant Information Matters: Filtering Out Noisy Samples to Boost RL

Résumé

In reinforcement learning, policy gradient algorithms optimize the policy directly and rely on sampling efficiently an environment. Nevertheless, while most sampling procedures are based on direct policy sampling, self-performance measures could be used to improve such sampling prior to each policy update. Following this line of thought, we introduce SAUNA, a method where non-informative transitions are rejected from the gradient update. The level of information is estimated according to the fraction of variance explained by the value function: a measure of the discrepancy between V and the empirical returns. In this work, we use this metric to select samples that are useful to learn from, and we demonstrate that this selection can significantly improve the performance of policy gradient methods. In this paper: (a) We define SAUNA's metric and introduce its method to filter transitions. (b) We conduct experiments on a set of benchmark continuous control problems. SAUNA significantly improves performance. (c) We investigate how SAUNA reliably selects samples with the most positive impact on learning and study its improvement on both performance and sample efficiency.
Fichier principal
Vignette du fichier
sauna_ijcai_camera_ready.pdf (1.75 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02091547 , version 1 (08-04-2019)
hal-02091547 , version 2 (10-04-2019)
hal-02091547 , version 3 (24-09-2019)
hal-02091547 , version 4 (25-09-2019)
hal-02091547 , version 5 (13-05-2020)
hal-02091547 , version 6 (20-11-2020)

Identifiants

Citer

Yannis Flet-Berliac, Philippe Preux. Only Relevant Information Matters: Filtering Out Noisy Samples to Boost RL. IJCAI 2020 - International Joint Conference on Artificial Intelligence, Jul 2020, Yokohama, Japan. ⟨10.24963/ijcai.2020/376⟩. ⟨hal-02091547v6⟩
333 Consultations
690 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More