arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.12914cs.CLcs.HC
收藏

基于大语言模型的人机协作求解复杂任务

Large Language Model-based Human-Agent Collaboration for Complex Task Solving

Gaoling School of Artificial Intelligence, Renmin University of China(中国人民大学高瓴人工智能学院);Department of Computer Science and Technology, Tsinghua University(清华大学计算机科学与技术系)

Xueyang Feng, Zhi-Yuan Chen, Yujia Qin, Yankai Lin, Xu Chen, Zhiyuan Liu, Ji-Rong Wen

AI总结:

本文提出ReHAC方法,通过强化学习策略模型确定人机协作中最佳人类干预时机,并构建离线训练数据集,验证了有限且合理的人类干预能显著提升基于LLM的智能体在复杂任务中的表现。

摘要:

在研究界近期的进展中,将大语言模型(LLMs)集成到完全自主智能体的创建中引起了广泛关注。尽管如此,基于LLM的智能体在适应动态环境和充分理解人类需求方面经常表现出明显的不足。在这项工作中,我们提出了基于LLM的人机协作求解复杂任务这一问题,探索二者的协同潜力。此外,我们提出了一种基于强化学习的人机协作方法ReHAC。该方法包含一个策略模型,旨在确定任务求解过程中最适合人类干预的阶段。我们构建了一个人机协作数据集,用于在离线强化学习环境中训练该策略模型。我们的验证测试证实了该模型的有效性。结果表明,人类与基于LLM的智能体的协同努力显著提升了复杂任务中的性能,这主要得益于经过良好规划的、有限的人类干预。数据集和代码可在以下地址获取:https://github.com/XueyangFeng/ReHAC。

英文摘要:

In recent developments within the research community, the integration of Large Language Models (LLMs) in creating fully autonomous agents has garnered significant interest. Despite this, LLM-based agents frequently demonstrate notable shortcomings in adjusting to dynamic environments and fully grasping human needs. In this work, we introduce the problem of LLM-based human-agent collaboration for complex task-solving, exploring their synergistic potential. In addition, we propose a Reinforcement Learning-based Human-Agent Collaboration method, ReHAC. This approach includes a policy model designed to determine the most opportune stages for human intervention within the task-solving process. We construct a human-agent collaboration dataset to train this policy model in an offline reinforcement learning environment. Our validation tests confirm the model's effectiveness. The results demonstrate that the synergistic efforts of humans and LLM-based agents significantly improve performance in complex tasks, primarily through well-planned, limited human intervention. Datasets and code are available at: https://github.com/XueyangFeng/ReHAC.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑