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

TL;DR

This paper tackles the challenge of factually accurate responses in knowledge-rich tasks by introducing SMART, a four-agent framework that leverages external knowledge through a Long-Short Trajectory Learning paradigm. The architecture splits tasks into Intent Reconstructor, Knowledge Retriever, Fact Locator, and Response Generator, and trains them via short-trajectory pretraining followed by long-trajectory synergy to model inter-agent flow with trajectory tokens. Empirical results across five knowledge-intensive benchmarks show SMART surpassing both knowledge-internalization and knowledge-enhancement baselines, including some larger LLMs, and demonstrate data-efficiency with significantly reduced long-trajectory data requirements. The work presents a general, plug-in framework for multi-agent systems to internalize tailored trajectories, with potential extensions to iterative optimization and retriever training for even more complex tasks.

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 https://github.com/yueshengbin/SMART.

Figures (7)

• Figure 1: Example of our long trajectory for knowledge-intensive scenarios (Top) and optimization comparison of multi-agent frameworks (Bottom). Solid and dashed arrows indicate inference and optimization paths, respectively.
• Figure 2: Overview of our multi-agent framework with long- and short-trajectory learning. This framework incorporates four agents: intent reconstructor, knowledge retriever, fac locator, and response generator.
• Figure 3: Overview of Long-Short Trajectory Learning. It consists of two stages, for short trajectory learning, under a given trajectory head, requires insight into the various explicit and implicit signals in each particular task. For long-trajectory learning, LLM executes the entire process by predicting different trajectory tokens, ensuring the synergism of different short-trajectories.
• Figure 4: Effects of long-trajectory training data size (K) on three tasks, ARC-C, PopQA and ASQA.
• Figure 5: Example of our SMART output on ARC-Challenge