迭代检索让模型答对更难的科学多跳问题，比直接给答案还强。
When Iterative RAG Beats Ideal Evidence: A Diagnostic Study in Scientific Multi-hop Question Answering
- 用循环检索+推理+动态停止，分步逼近答案
- 相比理想答案输入，准确率最高提升25.6个百分点
- 适合科研问答场景，尤其非微调模型
检索增强生成（RAG）扩展了大语言模型（LLM）的非参数知识能力，但其在需要多跳推理的科学领域中，迭代检索-推理循环是否优于静态检索仍不明确。本文首次进行机制级诊断评估，对比三种模式：无上下文（仅依赖参数记忆）、黄金上下文（一次性提供全部理想证据）和迭代RAG（无需训练的控制器，交替执行检索、假设优化和证据感知终止）。基于化学领域的ChemKGMultiHopQA数据集，分析检索覆盖率缺口、锚点丢失、查询质量、组合保真度与控制校准。结果表明，迭代RAG始终优于黄金上下文，尤其在非推理微调模型上表现突出，最高提升25.6个百分点。分阶段检索可减少后期失败、缓解上下文过载，并动态修正早期假设偏差，但仍有不完整跳覆盖、干扰项锁定、提前终止误判及高组合失败等问题。整体而言，分步检索过程的影响常大于理想证据的存在。本文为科学领域RAG部署与诊断提供实用指导。代码与结果见https://github.com/Matroid1998/Iterative-rag。
原文摘要 · Abstract (English)
Retrieval-Augmented Generation (RAG) extends large language models (LLMs) beyond parametric knowledge, yet it is unclear when iterative retrieval-reasoning loops meaningfully outperform static RAG, particularly in scientific domains requiring multi-hop reasoning over sparse, heterogeneous evidence. We provide the first controlled, mechanism-level diagnostic evaluation of whether synchronized iterative retrieval and reasoning can surpass even an idealized static upper bound (Gold Context) RAG. We benchmark eleven state-of-the-art LLMs under three regimes: (i) No Context, measuring reliance on parametric memory; (ii) Gold Context, where all oracle evidence is supplied at once; and (iii) Iterative RAG, a training-free controller that alternates retrieval, hypothesis refinement, and evidence-aware stopping. Using the chemistry-focused ChemKGMultiHopQA dataset, we isolate questions requiring genuine retrieval and analyze retrieval coverage gaps, anchor carry drop, query quality, composition fidelity, and control calibration. Iterative RAG consistently outperforms Gold Context, with gains up to 25.6 percentage points, especially for non-reasoning fine-tuned models. Staged retrieval reduces late-hop failures, mitigates context overload, and enables dynamic correction of early hypothesis drift, but failure modes remain, including incomplete hop coverage, distractor latch trajectories, early stopping miscalibration, and high composition failure rates even with perfect retrieval. Overall, the process of staged retrieval is often more influential than the mere presence of ideal evidence. We provide practical guidance for deploying and diagnosing RAG in specialized scientific settings. Code and evaluation results are available at https://github.com/Matroid1998/Iterative-rag
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。