RICE: Breaking Through the Training Bottlenecks of Reinforcement Learning with Explanation

TL;DR

This paper introduces RICE, a novel reinforcement learning refinement method that uses explanation techniques to create a better initial state distribution, helping agents overcome training bottlenecks and improve performance in complex tasks.

Contribution

RICE is a new refining scheme that integrates explanation methods to construct a mixed initial state distribution, providing theoretical guarantees and improved performance over existing methods.

Findings

RICE outperforms existing schemes in various RL environments.

The method effectively helps agents escape training bottlenecks.

Theoretical analysis shows tighter sub-optimality bounds.

Abstract

Deep reinforcement learning (DRL) is playing an increasingly important role in real-world applications. However, obtaining an optimally performing DRL agent for complex tasks, especially with sparse rewards, remains a significant challenge. The training of a DRL agent can be often trapped in a bottleneck without further progress. In this paper, we propose RICE, an innovative refining scheme for reinforcement learning that incorporates explanation methods to break through the training bottlenecks. The high-level idea of RICE is to construct a new initial state distribution that combines both the default initial states and critical states identified through explanation methods, thereby encouraging the agent to explore from the mixed initial states. Through careful design, we can theoretically guarantee that our refining scheme has a tighter sub-optimality bound. We evaluate RICE in various popular RL environments and real-world applications. The results demonstrate that RICE significantly outperforms existing refining schemes in enhancing agent performance.