RAG / 检索增强生成 - arXivDaily 专题

2504.11544 2025-04-17 cs.AI 83%

NodeRAG: Structuring Graph-based RAG with Heterogeneous Nodes

Tianyang Xu, Haojie Zheng, Chengze Li, Haoxiang Chen, Yixin Liu, Ruoxi Chen, Lichao Sun

专题命中长文档RAG ：RAG(title,abstract)；retrieval-augmented generation(abstract)；分类 cs.AI

2502.09017 2025-04-09 cs.CL cs.LG 83%

Diversity Enhances an LLM's Performance in RAG and Long-context Task

Zhichao Wang, Bin Bi, Yanqi Luo, Sitaram Asur, Claire Na Cheng

专题命中长文档RAG ：RAG(title,abstract)；retrieval-augmented generation(abstract)；分类 cs.CL

2501.01880 2025-01-06 cs.CL 83%

Long Context vs. RAG for LLMs: An Evaluation and Revisits

Xinze Li, Yixin Cao, Yubo Ma, Aixin Sun

专题命中长文档RAG ：RAG(title,abstract)；retrieval-augmented generation(abstract)；分类 cs.CL

Comments 14 pages excluding references and appendix

2404.02103 2024-12-24 cs.CL 83%

CLAPNQ: Cohesive Long-form Answers from Passages in Natural Questions for RAG systems

Sara Rosenthal, Avirup Sil, Radu Florian, Salim Roukos

专题命中长文档RAG ：RAG(title,abstract)；retrieval augmented generation(abstract)；分类 cs.CL

Comments 26 pages, Accepted at TACL

2411.19229 2024-12-17 cs.HC cs.AI cs.CY 83%

Habit Coach: Customising RAG-based chatbots to support behavior change

Arian Fooroogh Mand Arabi, Cansu Koyuturk, Michael O'Mahony, Raffaella Calati, Dimitri Ognibene

专题命中长文档RAG ：RAG(title,abstract)；retrieval-augmented generation(abstract)；分类 cs.AI

Comments Accepted for Italian Workshop on Artificial Intelligence for Human Machine Interaction (AIxHMI 2024), November 26, 2024, Bolzano, Italy

2411.03538 2024-11-07 cs.LG cs.CL 83%

Long Context RAG Performance of Large Language Models

Quinn Leng, Jacob Portes, Sam Havens, Matei Zaharia, Michael Carbin

专题命中长文档RAG ：RAG(title,abstract)；retrieval augmented generation(abstract)；分类 cs.CL

Comments 2024 NeurIPS workshop on Adaptive Foundation Models: Evolving AI for Personalized and Efficient Learning

2409.01666 2024-09-04 cs.CL 83%

In Defense of RAG in the Era of Long-Context Language Models

Tan Yu, Anbang Xu, Rama Akkiraju

专题命中长文档RAG ：RAG(title,abstract)；retrieval-augmented generation(abstract)；分类 cs.CL

2407.12325 2024-07-18 cs.IR 83%

Optimizing Query Generation for Enhanced Document Retrieval in RAG

Hamin Koo, Minseon Kim, Sung Ju Hwang

专题命中长文档RAG ：RAG(title,abstract)；retrieval-augmented generation(abstract)；分类 cs.IR

2405.03085 2024-05-07 cs.CL 83%

Compressing Long Context for Enhancing RAG with AMR-based Concept Distillation

Kaize Shi, Xueyao Sun, Qing Li, Guandong Xu

专题命中长文档RAG ：RAG(title,abstract)；retrieval augmented generation(abstract)；分类 cs.CL

2606.17255 2026-06-17 cs.CL cs.AI 新提交 82%

MLLP-VRAIN UPV system for the IWSLT 2026 Simultaneous Speech Translation task

MLLP-VRAIN UPV 系统在 IWSLT 2026 同声传译任务中的应用

Jorge Iranzo-Sánchez, Gerard Mas-Mollà, Adrià Giménez, Jorge Civera, Albert Sanchis, Alfons Juan

机构 * MLLP-VRAIN research group（MLLP-VRAIN研究组）； VRAIN ； Universitat Politècnica de València（瓦伦西亚理工大学）

专题命中长文档RAG ：RAG(summary_cn,abstract)；分类 cs.CL、cs.AI

AI总结提出基于Parakeet和Qwen 3.5模型的级联同声传译系统，通过自适应黑盒策略优化质量-延迟权衡，并引入ASR词增强和RAG机制处理上下文跟踪，在MCIF En→De测试集上实现XCOMET-XL提升+5.82。

Comments IWSLT 2026 System Description

2605.28742 2026-06-08 cs.AI 版本更新 81%

CORE: Contrastive Reflection Enables Rapid Improvements in Reasoning

CORE: 对比反思实现推理能力的快速提升

Linas Nasvytis, Simon Jerome Han, Ben Prystawski, Satchel Grant, Noah D. Goodman, Judith E. Fan

机构 * Stanford University（斯坦福大学）

专题命中长文档RAG ：RAG(summary_cn,abstract)；分类 cs.AI

AI总结提出对比反思（CORE）非参数学习算法，通过对比成功与失败的推理轨迹生成自然语言洞察，在少量样本和 rollout 下实现比参数方法（GRPO）和非参数方法（GEPA、情景RAG、MemRL）更快的推理性能提升。

2604.08585 2026-04-13 cs.DB cs.AI 81%

QCFuse: Query-Centric Cache Fusion for Efficient RAG Inference

QCFuse：基于查询的缓存融合用于高效RAG推理

Jianxin Yan, Zeheng Qian, Wangze Ni, Zhitao Shen, Zhiping Wang, Haoyang Li, Jia Zhu, Lei Chen, Kui Ren

机构 * Zhejiang University（浙江大学）； The University of Sydney（悉尼大学）； Zhejiang Normal University（浙江师范大学）

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.AI、cs.DB

AI总结 QCFuse通过基于查询的缓存融合提升RAG推理效率，利用语义摘要锚点优化查询表示并选择性重计算相关token，实验证明在保持准确性的同时提升响应效率40%。

2603.14045 2026-03-19 cs.IR cs.CL 81%

The Reasoning Bottleneck in Graph-RAG: Structured Prompting and Context Compression for Multi-Hop QA

图式推理瓶颈：用于多跳问答的结构化提示与上下文压缩

Yasaman Zarrinkia, Venkatesh Srinivasan, Alex Thomo

机构 * University of Victoria（维多利亚大学）； Santa Clara University（圣克拉拉大学）

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.IR、cs.CL

AI总结图式推理瓶颈：本文提出结构化提示和上下文压缩方法，通过SPARQL链式思考提示和图遍历压缩提升多跳问答准确性，验证其在不同Graph-RAG系统中的有效性。

Comments 11 pages, 2 figures, 9 tables; under review

2602.10271 2026-02-16 cs.IR cs.CL 81%

MLDocRAG: Multimodal Long-Context Document Retrieval Augmented Generation

MLDocRAG: 多模态长上下文文档检索增强生成

Yongyue Zhang, Yaxiong Wu

机构 * Independent Researcher（独立研究者）

专题命中长文档RAG ：retrieval augmented generation(title,abstract)；分类 cs.IR、cs.CL

AI总结 MLDocRAG通过构建多模态片段-查询图，实现多模态长上下文文档的检索增强生成，提升问答的准确性和连贯性。

Comments 15 pages

2601.08851 2026-01-15 cs.CL cs.AI 81%

Más contexto no es mejor. Paradoja de la dilución vectorial en RAG corporativos

更多上下文并不更好。RAG企业中的向量稀释悖论

Alex Dantart

机构 * Humanizing Internet（人类化互联网）

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.CL、cs.AI

AI总结该研究揭示了RAG中注入摘要导致向量稀释的问题，通过实验发现适度注入提升召回率但过度注入会降低精度，提出理论框架优化注入比例。

Comments in Spanish and English languages

2510.11956 2026-01-15 cs.CL cs.IR 81%

Investigating Retrieval-Augmented Generation Systems on Unanswerable, Uncheatable, Realistic, Multi-hop Queries

研究检索增强生成系统在无法回答、无法欺骗、现实、多跳查询上的应用

Gabrielle Kaili-May Liu, Bryan Li, Arman Cohan, William Gantt Walden, Eugene Yang

机构 * Yale University（耶鲁大学）； University of Pennsylvania（宾夕法尼亚大学）； Human Language Technology Center of Excellence（语言技术中心卓越中心）； Johns Hopkins University（约翰霍普金斯大学）

专题命中长文档RAG ：retrieval-augmented generation(title)；RAG(abstract)；分类 cs.IR、cs.CL

AI总结本文提出了一种自动创建难以欺骗、现实、无法回答和多跳查询的流程，以更真实地测试RAG系统的性能。

Comments ECIR 2026

2510.03521 2025-10-07 cs.CL cs.AI 81%

Identifying Financial Risk Information Using RAG with a Contrastive Insight

Ali Elahi

机构 * Department of Computer Science University of Illinois Chicago（伊利诺伊大学芝加哥分校计算机科学系）； Surlamer Investments（Surlamer投资公司）

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.CL、cs.AI

Comments 7 pages, 1 figure, Workshop on Generative AI in Finance, NeurIPS 2025

2508.04390 2025-08-07 cs.CL cs.AI 81%

AIC CTU@FEVER 8: On-premise fact checking through long context RAG

Herbert Ullrich, Jan Drchal

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.CL、cs.AI

2504.08756 2025-06-02 cs.IR cs.AI 81%

MHTS: Multi-Hop Tree Structure Framework for Generating Difficulty-Controllable QA Datasets for RAG Evaluation

Jeongsoo Lee, Daeyong Kwon, Kyohoon Jin, Junnyeong Jeong, Minwoo Sim, Minwoo Kim

机构 * DATUMO ； Graduate School of Culture Technology, Korea Advanced Institute of Science & Technology（韩国科学技术院文化科技研究生院）

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.IR、cs.AI

2410.12886 2024-10-18 cs.CL cs.IR cs.LG 81%

AT-RAG: An Adaptive RAG Model Enhancing Query Efficiency with Topic Filtering and Iterative Reasoning

Mohammad Reza Rezaei, Maziar Hafezi, Amit Satpathy, Lovell Hodge, Ebrahim Pourjafari

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.IR、cs.CL

2409.12294 2024-09-20 cs.AI cs.CL cs.LG cs.RO 81%

RAG-Modulo: Solving Sequential Tasks using Experience, Critics, and Language Models

Abhinav Jain, Chris Jermaine, Vaibhav Unhelkar

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.CL、cs.AI

Comments 8 pages, 5 figures

2506.17798 2026-08-11 cs.SE cs.CR 版本更新 80%

CognixShield: PoV-Guided Vulnerable API Usage Detection in Large Codebases via LLMs

CognixShield：基于LLM的、面向大型代码库的PoV引导式漏洞API使用检测

Quanzhi Fu, Wang Lingxiang, Wenjia Song, Gelei Deng, Yi Liu, Dan Williams, Ying Zhang

专题命中长文档RAG ：RAG(summary_cn,abstract)

AI总结 CognixShield是一款LLM驱动的框架，通过保留语义的AST碎片化、感知漏洞的多智能体RAG、PoV引导的语义推理三个核心组件，在57个Java应用上实现了优于现有工具的漏洞API使用检测性能。

Comments accepted in ESEM 2026

2606.25550 2026-06-25 cs.SE 新提交 80%

On the Viability of Requirements Generation From Code: An Experience Report

从代码生成需求的可行性：经验报告

Alexander Korn, Jone Bartel, Max Unterbusch, Andreas Vogelsang

专题命中长文档RAG ：RAG(summary_cn,abstract)

AI总结本文通过实验评估基于LLM和RAG的代理方法从源代码生成需求，发现LLM无法可靠生成未实现需求、高质量需求或引入需求气味，且单一人工难以检测气味，表明该方法尚不可行。

Comments Accepted for publication at the 13th International Workshop on Artificial Intelligence and Requirements Engineering (AIRE'26)

2604.18257 2026-04-21 cs.IR cs.AI cs.CL 80%

DocQAC: Adaptive Trie-Guided Decoding for Effective In-Document Query Auto-Completion

DocQAC：适应性前缀引导解码用于高效的文档内查询自动补全

Rahul Mehta, Kavin R, Indrajit Pal, Tushar Abhishek, Pawan Goyal, Manish Gupta

机构 * Microsoft Corporation（微软公司）； Indian Institute of Technology（印度理工学院）； Indian Institute of Technology Kharagpur（印度理工学院卡拉格浦）

专题命中长文档RAG ：RAG(abstract,abstract_cn)；retrieval-augmented generation(abstract)；分类 cs.IR、cs.CL、cs.AI

AI总结 DocQAC通过适应性前缀引导解码框架提升长文档内的搜索效率，利用文档特定上下文增强查询精度，优于现有基线模型。

2412.02835 2024-12-05 cs.CL cs.AI cs.IR cs.LG 80%

CAISSON: Concept-Augmented Inference Suite of Self-Organizing Neural Networks

Igor Halperin

专题命中长文档RAG ：retrieval-augmented generation(abstract)；RAG(abstract)；vector search(abstract)；分类 cs.IR、cs.CL、cs.AI

Comments 26 pages, 7 figures, 8 tables

2603.29085 2026-04-01 cs.AI 79%

PAR$^2$-RAG: Planned Active Retrieval and Reasoning for Multi-Hop Question Answering

PAR$^2$-RAG：计划主动检索与推理用于多跳问答

Xingyu Li, Rongguang Wang, Yuying Wang, Mengqing Guo, Chenyang Li, Tao Sheng, Sujith Ravi, Dan Roth

机构 * Oracle AI（甲骨文人工智能）

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.AI

AI总结 PAR$^2$-RAG通过分阶段的覆盖与承诺分离，结合广度优先锚定和深度优先细化，提升多跳问答的准确率和检索效果。

Comments 11 pages, 2 figures

2509.09467 2025-09-12 cs.AI 79%

Inteligencia Artificial jurídica y el desafío de la veracidad: análisis de alucinaciones, optimización de RAG y principios para una integración responsable

Alex Dantart

机构 * CIO LittleJohn（LittleJohn公司）

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.AI

Comments in Spanish and English languages

2504.12143 2025-07-28 cs.CR cs.AI 79%

ARCeR: an Agentic RAG for the Automated Definition of Cyber Ranges

Matteo Lupinacci, Francesco Blefari, Francesco Romeo, Francesco Aurelio Pironti, Angelo Furfaro

机构 * University of Calabria（卡拉布里亚大学）； IMT School for Advenced Studies（IMT高级研究学院）

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.AI

2502.09304 2025-06-04 cs.IR 79%

KET-RAG: A Cost-Efficient Multi-Granular Indexing Framework for Graph-RAG

Yiqian Huang, Shiqi Zhang, Xiaokui Xiao

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.IR

2505.21419 2025-05-29 cs.AI cs.OS 79%

Diagnosing and Resolving Cloud Platform Instability with Multi-modal RAG LLMs

Yifan Wang, Kenneth P. Birman

机构 * Computer Science Department, Cornell University（康奈尔大学计算机科学系）

专题命中长文档RAG ：RAG(title,abstract)；分类 cs.AI

Comments Published in EuroMLSys2025

Journal ref 2025, Association for Computing Machinery, Proceedings of the 5th Workshop on Machine Learning and Systems, 139-147, 9, series EuroMLSys '25

1. 长文档RAG 383 篇

NodeRAG: Structuring Graph-based RAG with Heterogeneous Nodes

Diversity Enhances an LLM's Performance in RAG and Long-context Task

Long Context vs. RAG for LLMs: An Evaluation and Revisits

CLAPNQ: Cohesive Long-form Answers from Passages in Natural Questions for RAG systems

Habit Coach: Customising RAG-based chatbots to support behavior change

Long Context RAG Performance of Large Language Models

In Defense of RAG in the Era of Long-Context Language Models

Optimizing Query Generation for Enhanced Document Retrieval in RAG

Compressing Long Context for Enhancing RAG with AMR-based Concept Distillation

MLLP-VRAIN UPV system for the IWSLT 2026 Simultaneous Speech Translation task

CORE: Contrastive Reflection Enables Rapid Improvements in Reasoning

QCFuse: Query-Centric Cache Fusion for Efficient RAG Inference

The Reasoning Bottleneck in Graph-RAG: Structured Prompting and Context Compression for Multi-Hop QA

MLDocRAG: Multimodal Long-Context Document Retrieval Augmented Generation

Más contexto no es mejor. Paradoja de la dilución vectorial en RAG corporativos

Investigating Retrieval-Augmented Generation Systems on Unanswerable, Uncheatable, Realistic, Multi-hop Queries

Identifying Financial Risk Information Using RAG with a Contrastive Insight

AIC CTU@FEVER 8: On-premise fact checking through long context RAG

MHTS: Multi-Hop Tree Structure Framework for Generating Difficulty-Controllable QA Datasets for RAG Evaluation

AT-RAG: An Adaptive RAG Model Enhancing Query Efficiency with Topic Filtering and Iterative Reasoning

RAG-Modulo: Solving Sequential Tasks using Experience, Critics, and Language Models

CognixShield: PoV-Guided Vulnerable API Usage Detection in Large Codebases via LLMs

On the Viability of Requirements Generation From Code: An Experience Report

DocQAC: Adaptive Trie-Guided Decoding for Effective In-Document Query Auto-Completion

CAISSON: Concept-Augmented Inference Suite of Self-Organizing Neural Networks

PAR$^2$-RAG: Planned Active Retrieval and Reasoning for Multi-Hop Question Answering

Inteligencia Artificial jurídica y el desafío de la veracidad: análisis de alucinaciones, optimización de RAG y principios para una integración responsable

ARCeR: an Agentic RAG for the Automated Definition of Cyber Ranges

KET-RAG: A Cost-Efficient Multi-Granular Indexing Framework for Graph-RAG

Diagnosing and Resolving Cloud Platform Instability with Multi-modal RAG LLMs