arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.03442cs.CL
收藏

A-RAG:通过分层检索接口扩展代理检索增强生成

A-RAG: Scaling Agentic Retrieval-Augmented Generation via Hierarchical Retrieval Interfaces

University of Science and Technology of China(中国科学技术大学);Metastone Technology(metastone技术公司)

Mingxuan Du, Benfeng Xu, Chiwei Zhu, Shaohan Wang, Pengyu Wang, Xiaorui Wang, Zhendong Mao

AI总结:

A-RAG通过分层检索接口提升检索增强生成效果,有效利用模型能力并适应不同任务。

摘要:

前沿语言模型已经展示了强大的推理能力和长周期工具使用能力。然而,现有的RAG系统未能利用这些能力。它们仍然依赖于两种范式:(1)设计一个算法,一次性检索段落并将它们连接到模型的输入中,或者(2)预先定义一个工作流程并提示模型逐步执行。这两种范式都不允许模型参与检索决策,阻碍了模型改进时的高效扩展。在本文中,我们介绍了A-RAG,一种代理RAG框架,它直接将分层检索接口暴露给模型。A-RAG提供了三种检索工具:关键词搜索、语义搜索和片段阅读,使代理能够自适应地在多个粒度上搜索和检索信息。在多个开放领域问答基准测试中,A-RAG在可比较或更低的检索令牌数量下一致优于现有方法,证明了A-RAG有效利用了模型能力,并能够动态适应不同的RAG任务。我们进一步系统研究了A-RAG随着模型大小和测试时计算资源的变化而扩展的情况。我们还将发布我们的代码和评估套件以促进未来研究。代码和评估套件可在https://github.com/Ayanami0730/arag获取。

英文摘要:

Frontier language models have demonstrated strong reasoning and long-horizon tool-use capabilities. However, existing RAG systems fail to leverage these capabilities. They still rely on two paradigms: (1) designing an algorithm that retrieves passages in a single shot and concatenates them into the model's input, or (2) predefining a workflow and prompting the model to execute it step-by-step. Neither paradigm allows the model to participate in retrieval decisions, preventing efficient scaling with model improvements. In this paper, we introduce A-RAG, an Agentic RAG framework that exposes hierarchical retrieval interfaces directly to the model. A-RAG provides three retrieval tools: keyword search, semantic search, and chunk read, enabling the agent to adaptively search and retrieve information across multiple granularities. Experiments on multiple open-domain QA benchmarks show that A-RAG consistently outperforms existing approaches with comparable or lower retrieved tokens, demonstrating that A-RAG effectively leverages model capabilities and dynamically adapts to different RAG tasks. We further systematically study how A-RAG scales with model size and test-time compute. We will release our code and evaluation suite to facilitate future research. Code and evaluation suite are available at https://github.com/Ayanami0730/arag.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑