paper-with-me

홈 › Papers

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

2024-05-24 · Siddhant Bhambri, Amrita Bhattacharjee, Durgesh Kalwar, Lin Guan, Huan Liu, Subbarao Kambhampati

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.

📄 PDF Abstract BibTeX arXiv:2405.15194

Code (0)

등록된 구현이 없습니다.

Tasks

Language ModellingLarge Language ModelMinecraftQ-Learningreinforcement-learningReinforcement LearningReinforcement Learning (RL)

Methods 이 논문이 사용한 방법론

A2C A2C, or Advantage Actor Critic, is a synchronous version of the A3C policy gradient method. As an alternative to the asynchronous…
Q-Learning Q-Learning is an off-policy temporal difference control algorithm: $$Q\left(S\_{t}, A\_{t}\right) \leftarrow Q\left(S\_{t}, A\_{t}\right) + \alpha\left[R_{t+1} +…
Entropy Regularization 설명 없음
PPO Proximal Policy Optimization, or PPO, is a policy gradient method for reinforcement learning. The motivation was to have an algorithm with the data efficiency and reliable…

Similar Papers 제목 키워드 기반

Enhancing Q-Learning with Large Language Model Heuristics

2024-05-06 · Xiefeng Wu

Q-learning excels in learning from feedback within sequential decision-making tasks but often requires extensive sampling to achieve significant improvements. While reward shaping can enhance learning efficiency, non-pot…

Decision MakingLanguage ModelingLanguage ModellingLarge Language Model+4

Automatic Reward Shaping from Multi-Objective Human Heuristics

2025-12-17 · Yuqing Xie, Jiayu Chen, Wenhao Tang, Ya Zhang 외 arxiv

Designing effective reward functions remains a central challenge in reinforcement learning, especially in multi-objective environments. In this work, we propose Multi-Objective Reward Shaping with Exploration (MORSE), a …

Reinforcement Learning

MAESTRO: Multi-Agent Environment Shaping through Task and Reward Optimization

2025-11-24 · Boyuan Wu arxiv

Cooperative Multi-Agent Reinforcement Learning (MARL) faces two major design bottlenecks: crafting dense reward functions and constructing curricula that avoid local optima in high-dimensional, non-stationary environment…

Multi-agent Reinforcement Learning

A new Potential-Based Reward Shaping for Reinforcement Learning Agent

2019-02-17 · Babak Badnava, Mona Esmaeili, Nasser Mozayani, Payman Zarkesh-Ha

Potential-based reward shaping (PBRS) is a particular category of machine learning methods which aims to improve the learning speed of a reinforcement learning agent by extracting and utilizing extra knowledge while perf…

Atari Gamesreinforcement-learningReinforcement LearningReinforcement Learning (RL)+1

Reward Shaping Using Convolutional Neural Network

2022-10-30 · Hani Sami, Hadi Otrok, Jamal Bentahar, Azzam Mourad 외

In this paper, we propose Value Iteration Network for Reward Shaping (VIN-RS), a potential-based reward shaping mechanism using Convolutional Neural Network (CNN). The proposed VIN-RS embeds a CNN trained on computed lab…

MuJoCo