用笔记管理知识，让大模型更准地查资料、生成答案。
DeepNote: Note-Centric Deep Retrieval-Augmented Generation
- 以笔记为载体，动态决定何时查资料、怎么查。
- 在多个基准测试中表现优于基线10.2%至20.1%。
- 适合需要高准确率和深度知识推理的问答任务。
检索增强生成（RAG）通过引入外部知识缓解大语言模型在问答任务中的事实错误与幻觉问题。然而，现有自适应RAG方法依赖大模型判断检索时机并直接使用检索内容进行生成，难以真实反映信息需求，也未能充分挖掘检索知识。我们提出DeepNote，一种基于笔记的自适应RAG框架，实现对知识源的深度且稳健的探索。DeepNote利用笔记作为知识提炼与累积的载体，在深入探索阶段，通过笔记决定检索时机、生成检索查询，并迭代评估知识增长，最终选择最优笔记用于答案生成。大量实验与分析表明，DeepNote显著优于所有基线模型（提升10.2%至20.1%），具备获取高密度、高质量知识的能力。此外，通过直接偏好优化（DPO）可进一步提升性能。代码与数据已公开于https://github.com/thunlp/DeepNote。
原文摘要 · Abstract (English)
Retrieval-Augmented Generation (RAG) mitigates factual errors and hallucinations in Large Language Models (LLMs) for question-answering (QA) by incorporating external knowledge. However, existing adaptive RAG methods rely on LLMs to predict retrieval timing and directly use retrieved information for generation, often failing to reflect real information needs and fully leverage retrieved knowledge. We develop DeepNote, an adaptive RAG framework that achieves in-depth and robust exploration of knowledge sources through note-centric adaptive retrieval. DeepNote employs notes as carriers for refining and accumulating knowledge. During in-depth exploration, it uses these notes to determine retrieval timing, formulate retrieval queries, and iteratively assess knowledge growth, ultimately leveraging the best note for answer generation. Extensive experiments and analyses demonstrate that DeepNote significantly outperforms all baselines (+10.2% to +20.1%) and exhibits the ability to gather knowledge with both high density and quality. Additionally, DPO further improves the performance of DeepNote. The code and data are available at https://github.com/thunlp/DeepNote.
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。