CIIR@LiveRAG 2025：通过自训练优化多智能体检索增强生成
CIIR@LiveRAG 2025: Optimizing Multi-Agent Retrieval Augmented Generation through Self-Training
AI总结：
本研究提出多智能体RAG框架mRAG，通过带奖励引导轨迹采样的自训练优化智能体协作，在SIGIR 2025 LiveRAG竞赛的DataMorgana衍生数据集上优于传统RAG基线，验证了其在复杂真实RAG任务中的有效性。
摘要：
本文提出mRAG，这是一个多智能体检索增强生成（RAG）框架，由负责规划、搜索、推理、协调等子任务的专用智能体构成。我们的系统采用带奖励引导轨迹采样的自训练范式，优化智能体间协作并提升响应生成效果。在SIGIR 2025 LiveRAG竞赛的DataMorgana衍生数据集上评估，mRAG表现优于传统RAG基线。我们进一步分析了竞赛结果，并通过案例研究展示该框架的优势，证明其对复杂真实RAG任务的有效性。
英文摘要：
This paper presents mRAG, a multi-agent retrieval-augmented generation (RAG) framework composed of specialized agents for subtasks such as planning, searching, reasoning, and coordination. Our system uses a self-training paradigm with reward-guided trajectory sampling to optimize inter-agent collaboration and enhance response generation. Evaluated on DataMorgana-derived datasets during the SIGIR 2025 LiveRAG competition, mRAG outperforms conventional RAG baselines. We further analyze competition outcomes and showcase the framework's strengths with case studies, demonstrating its efficacy for complex, real-world RAG tasks.
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。
