Synergistic Multi-Agent Framework with Trajectory Learning for Knowledge-Intensive Tasks

View a PDF of the paper titled Synergistic Multi-Agent Framework with Trajectory Learning for Knowledge-Intensive Tasks, by Shengbin Yue and Siyuan Wang and Wei Chen and Xuanjing Huang and Zhongyu Wei

View PDF
HTML (experimental)

Abstract:Recent advancements in Large Language Models (LLMs) have led to significant breakthroughs in various natural language processing tasks. However, generating factually consistent responses in knowledge-intensive scenarios remains a challenge due to issues such as hallucination, difficulty in acquiring long-tailed knowledge, and limited memory expansion. This paper introduces SMART, a novel multi-agent framework that leverages external knowledge to enhance the interpretability and factual consistency of LLM-generated responses. SMART comprises four specialized agents, each performing a specific sub-trajectory action to navigate complex knowledge-intensive tasks. We propose a multi-agent co-training paradigm, Long-Short Trajectory Learning, which ensures synergistic collaboration among agents while maintaining fine-grained execution by each agent. Extensive experiments on five knowledge-intensive tasks demonstrate SMART’s superior performance compared to widely adopted knowledge internalization and knowledge enhancement methods. Our framework can extend beyond knowledge-intensive tasks to more complex scenarios. Our code is available at this https URL.