arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.02170cs.CLcs.AIcs.MA
收藏

面向任务导向的智能体协作的动态LLM驱动智能体网络

A Dynamic LLM-Powered Agent Network for Task-Oriented Agent Collaboration

Tsinghua University(清华大学);Georgia Institute of Technology(佐治亚理工学院);Institute for AI Industry Research (AIR), Tsinghua University(清华大学人工智能国际研究院);Jiangsu Collaborative Innovation Center for Language Competence(江苏省语言能力协同创新中心);Stanford University(斯坦福大学)

Zijun Liu, Yanzhe Zhang, Peng Li, Yang Liu, Diyi Yang

AI总结:

提出动态LLM智能体网络DyLAN,通过两阶段范式自动选择智能体并动态协作,在多任务中优于基线且在MMLU特定科目准确率提升25.0%。

摘要:

近期研究表明,多个大型语言模型(LLM)驱动的智能体进行协作是解决任务的一种有前景的方法。然而,当前方法受限于使用固定数量的智能体和静态通信结构。在本研究中,我们提出从候选智能体中自动选择一个团队,以面向不同任务和领域的动态通信结构进行协作。具体而言,我们构建了一个名为动态LLM驱动智能体网络($\ extbf{DyLAN}$)的框架,用于LLM驱动的智能体协作,运行两阶段范式:(1)团队优化和(2)任务求解。在第一阶段,我们利用一种基于名为$\ extit{Agent Importance Score}$的无监督度量的$\ extit{agent selection}$算法,根据智能体在初步试验中的贡献,实现针对给定任务选择最佳智能体。然后,在第二阶段,选定的智能体根据查询动态协作。在实验上,我们证明DyLAN在代码生成、决策制定、一般推理和算术推理任务中优于强基线,且计算成本适中。在MMLU的特定科目中,在团队优化阶段选择智能体团队使DyLAN的准确率提高了25.0%。

英文摘要:

Recent studies show that collaborating multiple large language model (LLM) powered agents is a promising way for task solving. However, current approaches are constrained by using a fixed number of agents and static communication structures. In this work, we propose automatically selecting a team of agents from candidates to collaborate in a dynamic communication structure toward different tasks and domains. Specifically, we build a framework named Dynamic LLM-Powered Agent Network ($\textbf{DyLAN}$) for LLM-powered agent collaboration, operating a two-stage paradigm: (1) Team Optimization and (2) Task Solving. During the first stage, we utilize an $\textit{agent selection}$ algorithm, based on an unsupervised metric called $\textit{Agent Importance Score}$, enabling the selection of best agents according to their contributions in a preliminary trial, oriented to the given task. Then, in the second stage, the selected agents collaborate dynamically according to the query. Empirically, we demonstrate that DyLAN outperforms strong baselines in code generation, decision-making, general reasoning, and arithmetic reasoning tasks with moderate computational cost. On specific subjects in MMLU, selecting a team of agents in the team optimization stage improves accuracy by up to 25.0% in DyLAN.

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。

↑