MA-RAG：基于协作思维链推理的多智能体检索增强生成
MA-RAG: Multi-Agent Retrieval-Augmented Generation via Collaborative Chain-of-Thought Reasoning
AI总结：
MA-RAG通过协调Planner、Step Definer、Extractor和QA等专用智能体进行协作思维链推理，将复杂QA任务分解为子任务逐步优化检索与合成，在多基准上超越现有RAG方法，确立了高效可靠的多智能体RAG新范式。
摘要：
我们提出了 MA-RAG，一个用于检索增强生成（RAG）的多智能体框架，旨在解决复杂信息检索任务中固有的模糊性与推理挑战。不同于依赖端到端微调或孤立组件增强的传统 RAG 方法，MA-RAG 协调一组协作的专用 AI 智能体：Planner、Step Definer、Extractor 和 QA Agents，每个智能体负责 RAG 流程中的一个特定阶段。通过将任务分解为查询消歧、证据提取和答案合成等子任务，并使智能体能够通过思维链提示交流中间推理，MA-RAG 在保持模块化可解释性的同时，逐步优化检索与合成。在包括 NQ、HotpotQA、2WikimQA 和 TriviaQA 在内的多跳与模糊 QA 基准上的大量实验表明，MA-RAG 在所有模型规模上均显著优于独立 LLM 和现有 RAG 方法。值得注意的是，即使是配备 MA-RAG 的小型 LLaMA3-8B 模型也超越了更大的独立 LLM，而更大的变体（LLaMA3-70B 和 GPT-4o-mini）在具有挑战性的多跳数据集上创下了新的最先进结果。消融研究表明，Planner 和 Extractor 智能体对于多跳推理至关重要，且高容量模型对于 QA 智能体有效合成答案尤为重要。除了通用领域 QA，MA-RAG 还能泛化到医学 QA 等专业领域，在无需任何领域特定微调的情况下，取得了与领域特定模型相媲美的性能。我们的结果突显了协作式模块化推理在检索增强系统中的有效性：MA-RAG 不仅提高了答案的准确性和鲁棒性，还提供了可解释的中间推理步骤，为高效可靠的多智能体 RAG 确立了新范式。
英文摘要：
We present MA-RAG, a Multi-Agent framework for Retrieval-Augmented Generation (RAG) that addresses the inherent ambiguities and reasoning challenges in complex information-seeking tasks. Unlike conventional RAG methods that rely on end-to-end fine-tuning or isolated component enhancements, MA-RAG orchestrates a collaborative set of specialized AI agents: Planner, Step Definer, Extractor, and QA Agents, each responsible for a distinct stage of the RAG pipeline. By decomposing tasks into subtasks such as query disambiguation, evidence extraction, and answer synthesis, and enabling agents to communicate intermediate reasoning via chain-of-thought prompting, MA-RAG progressively refines retrieval and synthesis while maintaining modular interpretability. Extensive experiments on multi-hop and ambiguous QA benchmarks, including NQ, HotpotQA, 2WikimQA, and TriviaQA, demonstrate that MA-RAG significantly outperforms standalone LLMs and existing RAG methods across all model scales. Notably, even a small LLaMA3-8B model equipped with MA-RAG surpasses larger standalone LLMs, while larger variants (LLaMA3-70B and GPT-4o-mini) set new state-of-the-art results on challenging multi-hop datasets. Ablation studies reveal that both the planner and extractor agents are critical for multi-hop reasoning, and that high-capacity models are especially important for the QA agent to synthesize answers effectively. Beyond general-domain QA, MA-RAG generalizes to specialized domains such as medical QA, achieving competitive performance against domain-specific models without any domain-specific fine-tuning. Our results highlight the effectiveness of collaborative, modular reasoning in retrieval-augmented systems: MA-RAG not only improves answer accuracy and robustness but also provides interpretable intermediate reasoning steps, establishing a new paradigm for efficient and reliable multi-agent RAG.
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。
