Extracting Heuristics from Large Language Models for Reward Shaping in Reinforcement Learning

TL;DR

This paper explores using Large Language Models to generate heuristics for reward shaping in reinforcement learning, significantly improving sample efficiency across various domains by leveraging LLMs with or without a verifier.

Contribution

It introduces a novel method of extracting heuristics from LLMs to construct reward shaping functions, enhancing RL sample efficiency in sparse reward environments.

Findings

LLM-generated heuristics improve RL sample efficiency

Verifiers help assess heuristic quality

Significant gains across multiple RL algorithms and domains

Abstract

Reinforcement Learning (RL) suffers from sample inefficiency in sparse reward domains, and the problem is further pronounced in case of stochastic transitions. To improve the sample efficiency, reward shaping is a well-studied approach to introduce intrinsic rewards that can help the RL agent converge to an optimal policy faster. However, designing a useful reward shaping function for all desirable states in the Markov Decision Process (MDP) is challenging, even for domain experts. Given that Large Language Models (LLMs) have demonstrated impressive performance across a magnitude of natural language tasks, we aim to answer the following question: `Can we obtain heuristics using LLMs for constructing a reward shaping function that can boost an RL agent's sample efficiency?' To this end, we aim to leverage off-the-shelf LLMs to generate a plan for an abstraction of the underlying MDP. We further use this LLM-generated plan as a heuristic to construct the reward shaping signal for the downstream RL agent. By characterizing the type of abstraction based on the MDP horizon length, we analyze the quality of heuristics when generated using an LLM, with and without a verifier in the loop. Our experiments across multiple domains with varying horizon length and number of sub-goals from the BabyAI environment suite, Household, Mario, and, Minecraft domain, show 1) the advantages and limitations of querying LLMs with and without a verifier to generate a reward shaping heuristic, and, 2) a significant improvement in the sample efficiency of PPO, A2C, and Q-learning when guided by the LLM-generated heuristics.