Open the Black Box: Step-based Policy Updates for Temporally-Correlated Episodic Reinforcement Learning

TL;DR

This paper introduces TCE, a novel episodic reinforcement learning algorithm that leverages step information for policy updates, improving data efficiency and trajectory smoothness by combining step-based and episodic RL advantages.

Contribution

TCE effectively utilizes step information in episodic policy updates, bridging the gap between step-based and episodic RL while maintaining data efficiency and trajectory smoothness.

Findings

TCE achieves performance comparable to recent ERL methods.

TCE maintains data efficiency similar to state-of-the-art step-based RL.

TCE produces smoother trajectories in reinforcement learning tasks.

Abstract

Current advancements in reinforcement learning (RL) have predominantly focused on learning step-based policies that generate actions for each perceived state. While these methods efficiently leverage step information from environmental interaction, they often ignore the temporal correlation between actions, resulting in inefficient exploration and unsmooth trajectories that are challenging to implement on real hardware. Episodic RL (ERL) seeks to overcome these challenges by exploring in parameters space that capture the correlation of actions. However, these approaches typically compromise data efficiency, as they treat trajectories as opaque \emph{black boxes}. In this work, we introduce a novel ERL algorithm, Temporally-Correlated Episodic RL (TCE), which effectively utilizes step information in episodic policy updates, opening the 'black box' in existing ERL methods while retaining the smooth and consistent exploration in parameter space. TCE synergistically combines the advantages of step-based and episodic RL, achieving comparable performance to recent ERL methods while maintaining data efficiency akin to state-of-the-art (SoTA) step-based RL.