Efficient Soft Actor-Critic with LLM-Based Action-Level Guidance for Continuous Control

TL;DR

GuidedSAC introduces LLM-based action guidance to enhance exploration and convergence speed in reinforcement learning, achieving superior sample efficiency and performance in various control tasks.

Contribution

The paper proposes GuidedSAC, integrating large language models as supervisors to improve exploration and convergence in Soft Actor-Critic algorithms.

Findings

GuidedSAC outperforms standard SAC and exploration methods in sample efficiency.

Theoretical analysis confirms convergence guarantees are maintained.

Effective in both discrete and continuous control environments.

Abstract

We present GuidedSAC, a novel reinforcement learning (RL) algorithm that facilitates efficient exploration in vast state-action spaces. GuidedSAC leverages large language models (LLMs) as intelligent supervisors that provide action-level guidance for the Soft Actor-Critic (SAC) algorithm. The LLM-based supervisor analyzes the most recent trajectory using state information and visual replays, offering action-level interventions that enable targeted exploration. Furthermore, we provide a theoretical analysis of GuidedSAC, proving that it preserves the convergence guarantees of SAC while improving convergence speed. Through experiments in both discrete and continuous control environments, including toy text tasks and complex MuJoCo benchmarks, we demonstrate that GuidedSAC consistently outperforms standard SAC and state-of-the-art exploration-enhanced variants (e.g., RND, ICM, and E3B) in terms of sample efficiency and final performance.