Efficient Soft Actor-Critic with LLM-Based Action-Level Guidance for Continuous Control
This addresses the challenge of exploration in vast state-action spaces for RL practitioners, offering a novel hybrid approach that is incremental in combining LLMs with existing RL algorithms.
The paper tackles the problem of inefficient exploration in reinforcement learning by introducing GuidedSAC, which uses LLMs to provide action-level guidance to the Soft Actor-Critic algorithm, resulting in improved sample efficiency and final performance over standard SAC and other exploration methods in control environments.
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.