Awesome Agentic RAG

一份精心策划的 Agentic RAG 资源列表，涵盖检索增强生成（RAG）从基础范式到前沿 Agentic RAG 的核心论文与开源项目。

中文综述参见：检索增强生成 (RAG) 技术综述 - Yue Shui

---

目录

• Survey (综述)
• RAG Foundations (RAG 基础)
• Advanced & Modular RAG (高级与模块化 RAG)
• Core Retrieval (核心检索技术)
• Agentic RAG (智能体 RAG)
• GraphRAG (图 RAG)
• Multimodal RAG (多模态 RAG)
• Evaluation (评估)
• Applications (应用)
• Contributing
• License

---

Survey (综述)

Date	Title	Paper	Code	Comment
2025-07	Towards Agentic RAG with Deep Reasoning: A Survey of RAG-Reasoning Systems in LLMs	Paper	-	Li et al. 统一 RAG 与推理，系统梳理检索-推理融合方法、数据集与前沿框架
2025-06	Reasoning RAG via System 1 or System 2: A Survey on Reasoning Agentic RAG for Industry Challenges	Paper	-	Liang et al. 将 Agentic RAG 分为 System 1（预定义推理）与 System 2（自主代理推理）
2025-03	Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning	Paper	-	用 RL 训练 LLM 在推理过程中自主搜索，引入 retrieved token masking 实现稳定 RL 训练
2025-02	检索增强生成 (RAG) 技术综述	Paper	-	Yue Shui. 系统梳理 RAG 基础知识、范式演进、核心技术与前沿进展
2025-02	A Survey on Multimodal Retrieval Augmented Generation	Paper	-	Abootorabi et al. 多模态 RAG 全面综述
2025-01	Agentic Retrieval-Augmented Generation: A Survey on Agentic RAG	Paper	-	Singh et al. Agentic RAG 系统综述
2024-08	Graph Retrieval-Augmented Generation: A Survey	Paper	-	Peng et al. GraphRAG 领域综述
2023-12	Retrieval-Augmented Generation for Large Language Models: A Survey	Paper	-	Gao et al. RAG 大规模综述
2023-09	Retrieving Multimodal Information for Augmented Generation: A Survey	Paper	-	Zhao et al. EMNLP 2023. 多模态 RAG 综述

---

RAG Foundations (RAG 基础)

Date	Title	Paper	Code	Comment
2024-07	RAG vs. Long-Context LLMs: A Comprehensive Study and Hybrid Approach	Paper	-	Li et al. EMNLP 2024. RAG 与长上下文 LLM 的全面对比与混合方法
2020-05	Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks	Paper	-	Lewis et al. NeurIPS 2020. 首次系统提出 RAG，结合参数化记忆与非参数化记忆
2020-02	REALM: Retrieval-Augmented Language Model Pre-Training	Paper	-	Guu et al. ICML 2020. 将检索增强引入语言模型预训练

---

Advanced & Modular RAG (高级与模块化 RAG)

预检索优化 (Pre-Retrieval)

Date	Title	Paper	Code	Comment
2023-05	Query Expansion by Prompting	Paper	-	Jagerman et al. 利用 prompt 进行查询扩展
2023-05	Query Rewriting for Retrieval-Augmented Large Language Models	Paper	-	Ma et al. 查询重写优化检索
2022-12	Precise Zero-Shot Dense Retrieval without Relevance Labels (HyDE)	Paper	-	Gao et al. 零样本稠密检索，使用假设文档嵌入

检索中优化 (In-Retrieval)

Date	Title	Paper	Code	Comment
2024-02	BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings	Paper	-	Chen et al. 多语言、多功能、多粒度嵌入模型

后检索优化 (Post-Retrieval)

Date	Title	Paper	Code	Comment
2023-10	LLMLingua: Compressing Prompts for Accelerated Inference	Paper	-	Jiang et al. 上下文压缩加速推理
2023-10	RECOMP: Improving Retrieval-Augmented LMs with Selective Context Compression	Paper	-	Xu et al. 选择性上下文压缩

模块化 RAG (Modular RAG)

Date	Title	Paper	Code	Comment
2024-03	Adaptive-RAG: Learning to Adapt Retrieval-Augmented LLMs via Question Complexity	Paper	-	Jeong et al. 按查询复杂度动态选择检索策略
2024-01	Corrective Retrieval Augmented Generation (CRAG)	Paper	-	Yan et al. 检索结果自我纠正机制
2024-01	RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval	Paper	-	Sarthi et al. 递归聚类 + 摘要构建树状索引
2023-10	Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection	Paper	-	Asai et al. ICLR 2024. LLM 生成时自主决策检索并自我批判
2023-05	Active Retrieval Augmented Generation (FLARE)	Paper	-	Jiang et al. 主动检索，根据生成置信度触发

---

Core Retrieval (核心检索技术)

Date	Title	Paper	Code	Comment
2020-02	Efficient and Robust Approximate Nearest Neighbor Search Using HNSW	Paper	-	Malkov & Yashunin. 分层可导航小世界图 ANN 索引
2019-02	FAISS: A Library for Efficient Similarity Search	Paper	Code	Johnson et al. Meta 开源高效向量相似度搜索库
2009-07	Reciprocal Rank Fusion (RRF)	Paper	-	Cormack et al. SIGIR 2009. 多排序列表融合算法
2009-04	BM25 (Okapi) — The Probabilistic Relevance Framework	Paper	-	Robertson & Zaragoza. 经典全文检索排序算法

---

Agentic RAG (智能体 RAG)

Agentic RAG 将自主代理集成到 RAG 流程中，赋予系统动态决策、规划、工具使用和多代理协作能力。

Agent Frameworks & Capabilities (代理框架与能力)

Date	Title	Paper	Code	Comment
2026-06	HMARS: A Hierarchical Multi-Agent Memory System for Long-Context Reasoning	Paper	-	Li et al. 层级多代理记忆系统，将长上下文视为受管理记忆而非扁平语料库
2023-08	AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation	Paper	Code	Wu et al. 多代理对话框架
2023-03	Self-Refine: Iterative Refinement with Self-Feedback	Paper	-	Madaan et al. 自我反馈迭代改进
2023-03	Reflexion: Language Agents with Verbal Reinforcement Learning	Paper	-	Shinn et al. 语言代理的口头强化学习
2023-02	Toolformer: Language Models Can Teach Themselves to Use Tools	Paper	-	Schick et al. LLM 自主学会使用外部工具

Frontier Methods (前沿方法)

Date	Title	Paper	Code	Comment
2026-07	KidnapRAG: A Black-Box Attack for Hijacking Reasoning in Agentic RAG	Paper	-	Choi et al. 首个针对 Agentic RAG 的黑盒投毒攻击，劫持多步推理链
2026-06	MIRROR: Novelty-Constrained Memory-Guided MCTS Red-Teaming for Agentic RAG	Paper	-	Singh et al. IJCNN 2026. 记忆引导 MCTS + 新颖性约束多模态 Agentic RAG 红队测试
2026-05	Verbal-R3: Verbal Reranker as the Missing Bridge between Retrieval and Reasoning	Paper	-	Park et al. ACL 2026 Main. 语言化重排序器连接检索与推理，复杂 QA SOTA
2026-05	SEMA-RAG: A Self-Evolving Multi-Agent RAG for Medical Reasoning	Paper	-	Huang et al. ACL 2026 Findings. 三专家代理（Interpreter/Explorer/Arbiter）自适应检索医疗推理
2026-05	GRASP: Graph Agentic Search over Propositions for Multi-hop QA	Paper	-	Jenkins et al. 层级实体-命题-段落图代理搜索，减少 40-50% token 用量
2026-05	Superintelligent Retrieval Agent (SIRA)	Paper	-	Yang et al. 将多轮代理检索压缩为单一语料判别动作，超越 Perplexity 代理
2026-05	LatentRAG: Latent Reasoning and Retrieval for Efficient Agentic RAG	Paper	-	Zheng & Worring. 将 Agentic RAG 移至连续隐空间，~90% 延迟降低与显式方法匹配
2026-05	CuSearch: Curriculum Rollout Sampling via Search Depth for Agentic RAG	Paper	-	Shen et al. 课程式 rollout 采样 RL 训练 Agentic RAG，+11.8 EM 点提升
2026-05	AutoSearch: Adaptive Search Depth for Efficient Agentic RAG via Reinforcement Learning	Paper	-	Sun et al. ACL 2026 Findings. RL 自适应决定搜索终止时机，平衡成本与精度
2026-05	AgenticRAG: Agentic Retrieval for Enterprise Knowledge Bases	Paper	-	Suresh et al. 企业搜索轻量级代理封装，BRIGHT 召回@1 提升 21.8pp
2025-12	Process vs. Outcome Reward: Which is Better for Agentic RAG Reinforcement Learning	Paper	-	Zhang et al. NeurIPS 2025. 过程级 vs 结果级奖励 RL 训练，结果级奖励更优
2025-08	End-to-End Agentic RAG System Training for Traceable Diagnostic Reasoning (Deep-DxSearch)	Paper	-	Zheng et al. 端到端训练 Agentic RAG，可追溯诊断推理与证据归因
2025-03	Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning	Paper	-	Jin et al. RL 训练 LLM 自主搜索
2025-03	R1-Searcher: Incentivizing the Search Capability in LLMs via Reinforcement Learning	Paper	-	Song et al. 纯 RL 教会 LLM 自主调用搜索
2025-01	Search-o1: Agentic Retrieval-Augmented Reasoning Models	Paper	-	Li et al. Agentic RAG 集成到推理模型，含 Reason-in-Documents 模块
2025-01	CoRAG: Chain-of-Retrieval Augmented Generation	Paper	-	Wang et al. 训练模型生成中间检索链
2024-12	GeAR: Graph Expansion Enhanced Retrieval	Paper	-	Shen et al. 图扩展增强检索器，处理多跳查询
2024-01	Agent-G: Agentic Graph Retrieval Framework	Paper	-	Lee et al. 基于代理的图检索框架，多代理动态分工

---

GraphRAG (图 RAG)

通过构建知识图谱捕捉实体间的复杂关系，支持全局性问题回答。

Date	Title	Paper	Code	Comment
2025-02	HippoRAG 2: From RAG to Human-Like Long-Term Memory	Paper	-	Gutierrez et al. 将 RAG 推向类人长期记忆系统
2024-04	From Local to Global: A Graph RAG Approach to Query-Focused Summarization	Paper	Code	Edge et al. Microsoft. LLM 提取实体/关系 → 社区检测 → 分层摘要 → Map-Reduce 全局答案生成
2024-02	G-Retriever: Retrieval-Augmented Generation for Textual Graph Understanding	Paper	-	He et al. 文本图理解与问答

---

Multimodal RAG (多模态 RAG)

将 RAG 扩展到文本、图像、音频、视频等多种模态。

Cross-Modal Retrieval (跨模态检索)

Date	Title	Paper	Code	Comment
2024-11	OmniSearch: Adaptive Multimodal Search	Paper	-	Li et al. 自适应多模态搜索
2024-07	ColPali: Efficient Document Retrieval with Vision Language Models	Paper	-	Faysse et al. 视觉语言模型文档检索
2021-03	CLIP: Learning Transferable Visual Models from Natural Language Supervision	Paper	Code	Radford et al. OpenAI. 图文跨模态对齐

Frontier (前沿)

Date	Title	Paper	Code	Comment
2025-12	MegaRAG: Multimodal Knowledge Graph RAG	Paper	-	Hsiao et al. 多模态知识图谱 RAG
2025-02	ViDoRAG: Visual Document Retrieval with Multi-Agent RAG	Paper	-	Wang et al. 视觉文档多智能体 RAG，GMM 混合检索 + 迭代工作流
2024-11	M3DocRAG: Multi-Modal Multi-Page Document RAG	Paper	-	Cho et al. 多页文档展平嵌入

---

Evaluation (评估)

Evaluation Frameworks (评估框架)

Date	Title	Paper	Code	Comment
2026-07	Bayesian Uncertainty Propagation for Agentic RAG Pipelines	Paper	-	Donaldson et al. ICMIAM 2026. 贝叶斯网络传播不确定性跨规划/评估/生成阶段
2026-05	AgenticRAGTracer: A Hop-Aware Benchmark for Diagnosing Multi-Step Retrieval Reasoning	Paper	-	You et al. ACL 2026 Findings. 首个 Agentic RAG 逐跳检索推理质量诊断基准
2026-05	Is Agentic RAG Worth It? An Experimental Comparison of RAG Approaches	Paper	-	Ferrazzi et al. ACL 2026 Industry. 朴素 RAG、增强 RAG 与 Agentic RAG 实证对比
2025-01	SafeRAG: Safety Benchmark for RAG	Paper	-	Liang et al. RAG 安全性基准
2024-04	ClashEval: Knowledge Conflict Evaluation	Paper	-	Wu et al. 知识冲突评估
2024-01	CRUD-RAG: Chinese RAG Benchmark	Paper	-	Lyu et al. 中文 RAG 基准
2023-11	ARES: An Automated Evaluation Framework for RAG	Paper	-	Saad-Falcon et al. LLM 合成数据自动评估
2023-09	RAGAS: Automated Evaluation of Retrieval Augmented Generation	Paper	Code	Es et al. 自动化 RAG 评估框架
2023-09	RGB: Benchmarking RAG with Noise, Negation, and Integration	Paper	-	Chen et al. 噪声、拒答与信息整合评测

Benchmarks (通用基准)

Date	Title	Paper	Code	Comment
2022-10	MTEB: Massive Text Embedding Benchmark	Paper	Code	Muennighoff et al. 文本嵌入模型大规模基准
2021-04	BEIR: A Heterogeneous Benchmark for Information Retrieval	Paper	-	Thakur et al. 零样本信息检索基准
2020-09	KILT: Knowledge Intensive Language Tasks Benchmark	Paper	-	Petroni et al. 知识密集型任务基准

---

Applications (应用)

Domain	Date	Title	Paper	Code	Comment
Robotics	2026-06	Agentic RAG-VLM: Affordance-Aware RAG with Self-Reflective Planning for Robotic Grasping	Paper	-	Chen et al. 层次化 RAG 连接 VLM 理解与机器人抓取 + 自我反思恢复循环
QA	2026-05	From RAG to Agentic RAG for Faithful Islamic Question Answering	Paper	-	Bhatia et al. ACL 2026 Findings. Agentic RAG 结构化工具调用提升宗教 QA 事实忠实性
Education	2026-06	Carolina Guide: A Multi-Agent RAG System with Institutional Guardrails	Paper	-	Torsion & Zhou. 模块化多 Agent RAG，98.9% 检索成功率，机构安全护栏
Enterprise	2026-06	CAMI: Cost-Aware Agent-Guided Multi-Indexing for Semantic Retrieval	Paper	-	Qidwai et al. ACM CAIS 2026. 代理引导成本感知多索引策略，+9.4% recall@10
Finance	2024-03	AlphaFin: Benchmarking Financial Analysis	Paper	-	Zhang et al. 金融分析基准
Recommendation	2024-03	CoRAL: Collaborative Retrieval-Augmented LLMs for Recommendation	Paper	-	Wu et al. 协同检索增强推荐
Medical	2024-02	REALM: Retrieval Augmented Large Medical Model	Paper	-	Zhu et al. 医疗领域检索增强
Science	2023-06	MolReGPT: Molecular and Reaction Generative Pre-trained Transformer	Paper	-	Li et al. 分子科学 GPT
Legal	2023-06	ChatLaw: Open-Source Legal Large Language Model	Paper	Code	Cui et al. 开源法律大模型
Dialogue	2022-08	BlenderBot 3: A Deployed Conversational Agent	Paper	Code	Shuster et al. Meta 部署级对话代理
Code	2022-07	DocPrompting: Generating Code by Retrieving the Docs	Paper	-	Zhou et al. 文档检索增强代码生成
QA	2017-04	Reading Wikipedia to Answer Open-Domain Questions (DrQA)	Paper	Code	Chen et al. 开放域问答经典工作

---

Contributing

欢迎贡献！请通过提交 Pull Request 或 Issue 来建议新的论文或资源。

快速贡献步骤

1. Fork 这个仓库
2. 创建你的特性分支 (git checkout -b add-new-resource)
3. 在表格中添加新资源（按时间倒序排列）
4. 提交你的改变 (git commit -m 'Add: [Resource Title]')
5. 推送到分支 (git push origin add-new-resource)
6. 开启一个 Pull Request

表格字段说明

• Date: 论文发表日期 (YYYY-MM 格式，arXiv 论文以 arXiv ID 中的年月为准)
• Title: 论文标题
• Paper: 论文链接 (arXiv / 会议 / 博客)
• Code: 开源代码链接 (如无则留空 -)
• Comment: 简要描述主要贡献，包含作者、会议/期刊信息

---

License

本项目采用 CC0 1.0 Universal License 许可证。