Adversarially Guided Actor-Critic - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year :

Adversarially Guided Actor-Critic

Yannis Flet-Berliac
Olivier Pietquin
Philippe Preux
Matthieu Geist

Abstract

Despite definite success in deep reinforcement learning problems, actor-critic algorithms are still confronted with sample inefficiency in complex environments, particularly in tasks where efficient exploration is a bottleneck. These methods consider a policy (the actor) and a value function (the critic) whose respective losses are built using different motivations and approaches. This paper introduces a third protagonist: the adversary. While the adversary mimics the actor by minimizing the KL-divergence between their respective action distributions, the actor, in addition to learning to solve the task, tries to differentiate itself from the adversary predictions. This novel objective stimulates the actor to follow strategies that could not have been correctly predicted from previous trajectories, making its behavior innovative in tasks where the reward is extremely rare. Our experimental analysis shows that the resulting Adversarially Guided Actor-Critic (AGAC) algorithm leads to more exhaustive exploration. Notably, AGAC outperforms current state-of-the-art methods on a set of various hard-exploration and procedurally-generated tasks.
Fichier principal
Vignette du fichier
AGAC.pdf (1.55 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03167169 , version 1 (11-03-2021)

Identifiers

Cite

Yannis Flet-Berliac, Johan Ferret, Olivier Pietquin, Philippe Preux, Matthieu Geist. Adversarially Guided Actor-Critic. ICLR 2021 - International Conference on Learning Representations, May 2021, Vienna / Virtual, Austria. ⟨hal-03167169⟩
63 View
186 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More