CIIR@LiveRAG 2025: Optimizing Multi-Agent Retrieval Augmented Generation through Self-Training
This work addresses complex, real-world RAG tasks, but it appears incremental as it builds on existing multi-agent and self-training paradigms without claiming a fundamental shift.
The paper tackles the problem of optimizing multi-agent retrieval-augmented generation (RAG) by introducing mRAG, a framework with specialized agents for planning, searching, reasoning, and coordination, which outperforms conventional RAG baselines on DataMorgana-derived datasets in the SIGIR 2025 LiveRAG competition.
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.