Deep Exploration with PAC-Bayes

Bahareh Tasdighi, Manuel Haussmann, Nicklas Werge, Yi-Shan Wu, Melih Kandemir

Keywords: Reinforcement learning, deep exploration, sparse rewards, PAC Bayes

TL;DR: We propose a novel actor-critic algorithm for continuous control in sparse reward settings, using a critic ensemble and PAC-Bayesian bounds on Bellman operator error to enable deep exploration via posterior sampling.

Abstract:

Reinforcement learning for continuous control under sparse rewards is an under-explored problem despite its significance in real life. Many complex skills build on intermediate ones as prerequisites. For instance, a humanoid locomotor has to learn how to stand before it can learn to walk. To cope with reward sparsity, a reinforcement learning agent has to perform deep exploration. However, existing deep exploration methods are designed for small discrete action spaces, and their successful generalization to state-of-the-art continuous control remains unproven. We address the deep exploration problem for the first time from a PAC-Bayesian perspective in the context of actor-critic learning. To do this, we quantify the error of the Bellman operator through a PAC-Bayes bound, where a bootstrapped ensemble of critic networks represents the posterior distribution, and their targets serve as a data-informed function-space prior.
We derive an objective function from this bound and use it to train the critic ensemble. Each critic trains an individual actor network, implemented as a shared trunk and critic-specific heads. The agent performs deep exploration by acting deterministically on a randomly chosen actor head. Our proposed algorithm, named PAC-Bayesian Actor-Critic (PBAC), is the only algorithm to successfully discover sparse rewards on a diverse set of continuous control tasks with varying difficulty.

Primary Area: reinforcement learning

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 6558