RAG / 检索增强生成

NanoVDR: Distilling a 2B Vision-Language Retriever into a 70M Text-Only Encoder for Visual Document Retrieval

NanoVDR：将20亿参数的视觉语言检索器蒸馏为7000万参数的纯文本编码器用于视觉文档检索

Zhuchenyang Liu, Yao Zhang, Yu Xiao

机构 * Aalto University（阿alto大学） 纠错

专题命中 检索器与排序 ：retriever(title)；分类 cs.IR

AI总结 利用查询-文档不对称性，通过蒸馏将20亿参数的视觉语言模型教师蒸馏为7000万参数的纯文本学生编码器，采用点态余弦对齐目标，实现视觉文档检索的高效推理。

STEM: Structure-Tracing Evidence Mining for Knowledge Graphs-Driven Retrieval-Augmented Generation

STEM: 用于知识图谱驱动检索增强生成的结构追踪证据挖掘

Peng Yu, En Xu, Bin Chen, Haibiao Chen, Yinfei Xu

机构 * AI Product Center, Kingsoft Corporation（金山办公人工智能产品中心） 纠错

专题命中 检索器与排序 ：retrieval-augmented generation(title)；分类 cs.CL

AI总结 本文提出STEM框架，通过将多跳推理重新定义为以模式引导的图搜索任务，解决知识图谱结构异质性和现有推理路径检索方法缺乏全局结构视角的问题，从而提升多跳推理的准确性和证据完整性。

Comments 34 pages, 16 figures, accepted to ACL 2026 (Main Conference, Oral Presentation)

MoshiRAG: Asynchronous Knowledge Retrieval for Full-Duplex Speech Language Models

MoshiRAG: 异步知识检索用于全双工语音语言模型

Chung-Ming Chien, Manu Orsini, Eugene Kharitonov, Neil Zeghidour, Karen Livescu, Alexandre Défossez

机构 * Toyota Technological Institute at Chicago（丰田技术研究所（芝加哥）） 纠错

专题命中 检索器与排序 ：knowledge retrieval(title)；分类 cs.CL

AI总结 本文提出MoshiRAG，通过异步框架结合紧凑接口与选择性检索，提升全双工语音模型的事实性，同时保持交互性。

Comments Accepted to ICML 2026

Loom: Hybrid Retrieval-Scoring Outfit Recommendation with Semantic Material Compatibility and Occasion-Aware Embedding Priors

Loom：基于语义材料兼容性和场合感知嵌入先验的混合检索-评分服装推荐

Anushree Berlia

专题命中 检索器与排序 ：hybrid retrieval(title)；分类 cs.IR

AI总结 Loom结合神经嵌入检索与结构化领域评分，通过语义材料权重和场合感知嵌入先验生成高质量服装搭配，实验表明其在评分和违规率上显著优于基线方法。

Comments Code: https://github.com/anushreeberlia/loom

Beyond Hard Negatives: The Importance of Score Distribution in Knowledge Distillation for Dense Retrieval

超越硬负样本：知识蒸馏中密集检索中分数分布的重要性

Youngjoon Jang, Seongtae Hong, Hyeonseok Moon, Heuiseok Lim

专题命中 检索器与排序 ：dense retrieval(title)；分类 cs.IR

AI总结 本文指出仅依赖硬负样本限制了学生模型学习教师偏好结构的能力，提出分层采样策略以均匀覆盖分数范围，实验表明该方法在不同设置中优于top-K和随机采样。

Comments SIGIR 2026

ARHN: Answer-Centric Relabeling of Hard Negatives with Open-Source LLMs for Dense Retrieval

ARHN: 基于开源大语言模型的密集检索中以答案为中心的困难负样本重标定

Hyewon Choi, Jooyoung Choi, Hansol Jang, Hyun Kim, Chulmin Yun, ChangWook Jun, Stanley Jungkyu Choi

专题命中 检索器与排序 ：dense retrieval(title)；分类 cs.IR

AI总结 ARHN通过利用开源大语言模型对困难负样本进行重标定，结合重标定与过滤策略提升检索效果，提供更清洁的监督信号。

Comments Accepted to SIGIR 2026

The Wisdom of Many Queries: Complexity-Diversity Principle for Dense Retriever Training

众查询的智慧：密集检索器训练的复杂度-多样性原则

Xincan Feng, Noriki Nishida, Yusuke Sakai, Yuji Matsumoto

专题命中 检索器与排序 ：retriever(title)；分类 cs.IR

AI总结 本文提出复杂度-多样性原则，通过多查询合成提升密集检索器在领域外任务中的泛化能力，发现查询复杂度与多样性间存在强正相关，提出基于复杂度的训练策略提升推理任务表现。

Comments Under review

Detecting Overflow in Compressed Token Representations for Retrieval-Augmented Generation

检测压缩标记表示中的溢出以用于检索增强生成

Julia Belikova, Danila Rozhevskii, Dennis Svirin, Konstantin Polev, Alexander Panchenko

机构 * Skoltech（斯克利切夫斯基技术学院） ； Sber AI Lab（Sber AI实验室） ； AIRI ； Institute for Information Transmission Problems of the Russian Academy of Sciences（俄罗斯科学院信息传输问题研究所） 纠错

专题命中 检索器与排序 ：retrieval-augmented generation(title)；分类 cs.CL

AI总结 本文提出了一种检测压缩标记表示中溢出的方法，通过轻量级探测分类器在多个数据集上实现了较高的检测性能，提升了检索增强生成任务的可靠性。

Comments Accepted to EACL 2026 Student Research Workshop. 14 pages, 6 tables, 1 figure

Latent Retrieval Augmented Generation of Cross-Domain Protein Binders

Zishen Zhang, Xiangzhe Kong, Wenbing Huang, Yang Liu

机构 * Dept. of Comp. Sci. & Tech., Tsinghua University（计算机科学与技术系，清华大学） ； Institute for AIR, Tsinghua University（空气智能研究所，清华大学） ； Gaoling School of Artificial Intelligence, Renmin University of China（北京理工大学人工智能学院） 纠错

专题命中 检索器与排序 ：retrieval augmented generation(title)；分类 cs.AI

Rowen: Adaptive Retrieval-Augmented Generation for Hallucination Mitigation in LLMs

Hanxing Ding, Liang Pang, Zihao Wei, Huawei Shen, Xueqi Cheng

机构 * State Key Laboratory of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences（人工智能安全国家重点实验室，计算技术研究所，中国科学院） 纠错

专题命中 检索器与排序 ：retrieval-augmented generation(title)；分类 cs.CL

Comments Accepted at SIGIR-AP 2025

Decoding Dense Embeddings: Sparse Autoencoders for Interpreting and Discretizing Dense Retrieval

Seongwan Park, Taeklim Kim, Youngjoong Ko

机构 * SungKyunKwan University（顺天妇女大学） 纠错

专题命中 检索器与排序 ：dense retrieval(title)；分类 cs.IR

Comments Published at EMNLP 2025 main

eSapiens: A Platform for Secure and Auditable Retrieval-Augmented Generation

Isaac Shi, Zeyuan Li, Fan Liu, Wenli Wang, Lewei He, Yang Yang, Tianyu Shi

机构 * eSapiens Team（eSapiens团队） 纠错

专题命中 检索器与排序 ：retrieval-augmented generation(title)；分类 cs.AI

Conventional Contrastive Learning Often Falls Short: Improving Dense Retrieval with Cross-Encoder Listwise Distillation and Synthetic Data

Manveer Singh Tamber, Suleman Kazi, Vivek Sourabh, Jimmy Lin

专题命中 检索器与排序 ：dense retrieval(title)；分类 cs.IR

Comments updated version of arxiv:2502.19712

Large Language Model as Universal Retriever in Industrial-Scale Recommender System

Junguang Jiang, Yanwen Huang, Bin Liu, Xiaoyu Kong, Xinhang Li, Ziru Xu, Han Zhu, Jian Xu, Bo Zheng

机构 * Taobao & Tmall Group of Alibaba, China（淘宝与天猫集团（阿里巴巴）） 纠错

专题命中 检索器与排序 ：retriever(title)；分类 cs.IR

CodeGRAG: Bridging the Gap between Natural Language and Programming Language via Graphical Retrieval Augmented Generation

Kounianhua Du, Jizheng Chen, Renting Rui, Huacan Chai, Lingyue Fu, Wei Xia, Yasheng Wang, Ruiming Tang, Yong Yu, Weinan Zhang

机构 * Shanghai Jiao Tong University（上海交通大学） ； Huawei Noah’s Ark Lab Shanghai（华为诺亚实验室） 纠错

专题命中 检索器与排序 ：retrieval augmented generation(title)；分类 cs.AI

Efficient Conversational Search via Topical Locality in Dense Retrieval

Cristina Ioana Muntean, Franco Maria Nardini, Raffaele Perego, Guido Rocchietti, Cosimo Rulli

专题命中 检索器与排序 ：dense retrieval(title)；分类 cs.IR

Comments 5 pages, 2 figures, SIGIR 2025

Boolean-aware Attention for Dense Retrieval

Quan Mai, Susan Gauch, Douglas Adams

专题命中 检索器与排序 ：dense retrieval(title)；分类 cs.CL

Momentum Posterior Regularization for Multi-hop Dense Retrieval

Zehua Xia, Yuyang Wu, Yiyun Xia, Cam-Tu Nguyen

专题命中 检索器与排序 ：dense retrieval(title)；分类 cs.CL

Comments Accepted by COLING 2025

Interpret and Control Dense Retrieval with Sparse Latent Features

Hao Kang, Tevin Wang, Chenyan Xiong

专题命中 检索器与排序 ：dense retrieval(title)；分类 cs.IR

Enhancing Knowledge Retrieval with In-Context Learning and Semantic Search through Generative AI

Mohammed-Khalil Ghali, Abdelrahman Farrag, Daehan Won, Yu Jin

专题命中 检索器与排序 ：knowledge retrieval(title)；分类 cs.IR

Optimizing Code Runtime Performance through Context-Aware Retrieval-Augmented Generation

Manish Acharya, Yifan Zhang, Kevin Leach, Yu Huang

专题命中 检索器与排序 ：retrieval-augmented generation(title)；分类 cs.AI

Real-time Indexing for Large-scale Recommendation by Streaming Vector Quantization Retriever

Xingyan Bin, Jianfei Cui, Wujie Yan, Zhichen Zhao, Xintian Han, Chongyang Yan, Feng Zhang, Xun Zhou, Qi Wu, Zuotao Liu

专题命中 检索器与排序 ：retriever(title)；分类 cs.IR

A Soft Sensor Method with Uncertainty-Awareness and Self-Explanation Based on Large Language Models Enhanced by Domain Knowledge Retrieval

Shuo Tong, Han Liu, Runyuan Guo, Wenqing Wang, Xueqiong Tian, Lingyun Wei, Lin Zhang, Huayong Wu, Ding Liu, Youmin Zhang

专题命中 检索器与排序 ：knowledge retrieval(title)；分类 cs.AI

ECLIPSE: Contrastive Dimension Importance Estimation with Pseudo-Irrelevance Feedback for Dense Retrieval

Giulio D'Erasmo, Giovanni Trappolini, Nicola Tonellotto, Fabrizio Silvestri

专题命中 检索器与排序 ：dense retrieval(title)；分类 cs.IR

DRUM: Learning Demonstration Retriever for Large MUlti-modal Models

Ellen Yi-Ge, Jiechao Gao, Wei Han, Wei Zhu

专题命中 检索器与排序 ：retriever(title)；分类 cs.CL

Inference is All You Need: Self Example Retriever for Cross-domain Dialogue State Tracking with ChatGPT

Jihyun Lee, Gary Geunbae Lee

专题命中 检索器与排序 ：retriever(title)；分类 cs.CL

Measuring text summarization factuality using atomic facts entailment metrics in the context of retrieval augmented generation

N. E. Kriman

专题命中 检索器与排序 ：retrieval augmented generation(title)；分类 cs.CL

Comments 12 pages

FLASH: Federated Learning-Based LLMs for Advanced Query Processing in Social Networks through RAG

Sai Puppala, Ismail Hossain, Md Jahangir Alam, Sajedul Talukder

专题命中 检索器与排序 ：RAG(title)；分类 cs.IR

Comments This research paper is submitted to ASONAM 2024 conference on Advances in Social Networks Analysis and Mining and going to be published in Springer

Accuracy and Consistency of LLMs in the Registered Dietitian Exam: The Impact of Prompt Engineering and Knowledge Retrieval

Iman Azimi, Mohan Qi, Li Wang, Amir M. Rahmani, Youlin Li

专题命中 检索器与排序 ：knowledge retrieval(title)；分类 cs.CL

Dense Retrieval with Continuous Explicit Feedback for Systematic Review Screening Prioritisation

Xinyu Mao, Shengyao Zhuang, Bevan Koopman, Guido Zuccon

专题命中 检索器与排序 ：dense retrieval(title)；分类 cs.IR

Comments Accepted at SIGIR 2024;typos corrected