Adaptive In-conversation Team Building for Language Model Agents

2024-05-29

arXiv.org

AI Summary

To address the challenge of dynamically adapting multi-LLM agent teams to diverse tasks, this paper proposes the Captain Agent framework—a dialogue-driven paradigm for real-time formation, reconfiguration, and orchestration of specialized agent teams without task-specific prompt engineering. Its core innovations include a nested group-chat mechanism, dynamic role assignment, a procedural reflection module, and collaborative scheduling strategies, jointly ensuring both agent diversity and system robustness. The framework enhances the capabilities of weaker LLMs through synergistic collaboration. Evaluated across six realistic scenarios, it achieves an average accuracy improvement of 21.94%, significantly elevating dialogue quality for low-capability models. Notably, it attains competitive performance at minimal computational and engineering cost, establishing— for the first time—the paradigm of adaptive, conversation-guided multi-agent team composition.

Technology Category

• Multiagent Systems: Teamwork
• Humans and AI: Teamwork, Team formation
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Leveraging multiple large language model (LLM) agents has shown to be a promising approach for tackling complex tasks, while the effective design of multiple agents for a particular application remains an art. It is thus intriguing to answer a critical question: Given a task, how can we build a team of LLM agents to solve it effectively? Our new adaptive team-building paradigm offers a flexible solution, realized through a novel agent design named Captain Agent. It dynamically forms and manages teams for each step of a task-solving process, utilizing nested group conversations and reflection to ensure diverse expertise and prevent stereotypical outputs, allowing for a flexible yet structured approach to problem-solving. A comprehensive evaluation across six real-world scenarios demonstrates that Captain Agent significantly outperforms existing multi-agent methods with 21.94% improvement in average accuracy, providing outstanding performance without requiring task-specific prompt engineering. Our exploration of different backbone LLM and cost analysis further shows that Captain Agent can improve the conversation quality of weak LLM and achieve competitive performance with extremely low cost, which illuminates the application of multi-agent systems.

Problem

• How to build effective LLM agent teams for complex tasks.
• Dynamic team formation and management using Captain Agent.
• Improving multi-agent performance without task-specific prompt engineering.

Innovation

• Dynamic team formation via Captain Agent
• Nested group conversations enhance expertise
• Low-cost, high-performance multi-agent system