LLM-Empowered State Representation for Reinforcement Learning
This addresses the challenge of low sample efficiency and high time costs in reinforcement learning for robotics and simulation tasks, representing an incremental advancement by integrating LLMs into state representation.
The paper tackles the problem of incomplete state representations in reinforcement learning by proposing LLM-Empowered State Representation (LESR), which uses large language models to generate task-related codes, resulting in an average 29% improvement in accumulated reward in Mujoco tasks and 30% in success rates in Gym-Robotics tasks.
Conventional state representations in reinforcement learning often omit critical task-related details, presenting a significant challenge for value networks in establishing accurate mappings from states to task rewards. Traditional methods typically depend on extensive sample learning to enrich state representations with task-specific information, which leads to low sample efficiency and high time costs. Recently, surging knowledgeable large language models (LLM) have provided promising substitutes for prior injection with minimal human intervention. Motivated by this, we propose LLM-Empowered State Representation (LESR), a novel approach that utilizes LLM to autonomously generate task-related state representation codes which help to enhance the continuity of network mappings and facilitate efficient training. Experimental results demonstrate LESR exhibits high sample efficiency and outperforms state-of-the-art baselines by an average of 29% in accumulated reward in Mujoco tasks and 30% in success rates in Gym-Robotics tasks.