Information-Directed Exploration for Deep Reinforcement Learning

Abstract

Efficient exploration remains a major challenge for reinforcement learning. One reason is that the variability of the returns often depends on the current state and action, and is therefore heteroscedastic. Classical exploration strategies such as upper confidence bound algorithms and Thompson sampling fail to appropriately account for heteroscedasticity, even in the bandit setting. Motivated by recent findings that address this issue in bandits, we propose to use Information-Directed Sampling (IDS) for exploration in reinforcement learning. As our main contribution, we build on recent advances in distributional reinforcement learning and propose a novel, tractable approximation of IDS for deep Q-learning. The resulting exploration strategy explicitly accounts for both parametric uncertainty and heteroscedastic observation noise. We evaluate our method on Atari games and demonstrate a significant improvement over alternative approaches.

Cited by in corpus (7)

• Estimating Risk and Uncertainty in Deep Reinforcement Learning
• DFAC Framework: Factorizing the Value Function via Quantile Mixture for Multi-Agent Distributional Q-Learning
• Principled Exploration via Optimistic Bootstrapping and Backward Induction
• Bridging Imagination and Reality for Model-Based Deep Reinforcement Learning
• Reinforcement Learning, Bit by Bit
• The Value of Information When Deciding What to Learn
• GMAC: A Distributional Perspective on Actor-Critic Framework