arXivDaily arXiv每日学术速递 周一至周五更新
​通知:联邦节假日调休:哥伦布日 / 原住民日(Columbus Day / Indigenous Peoples' Day),今日(13日)无更新。
arXiv 2506.23139cs.CLcs.AI

异构企业数据上的深度搜索基准测试

Benchmarking Deep Search over Heterogeneous Enterprise Data

Salesforce AI Research(Salesforce人工智能研究)

Prafulla Kumar Choubey, Xiangyu Peng, Shilpa Bhagavath, Kung-Hsiang Huang, Caiming Xiong, Chien-Sheng Wu

AI总结:

提出了一个评估深度搜索的新基准,利用合成数据管道模拟业务工作流生成多跳问题,发现现有智能体RAG方法因检索瓶颈导致性能显著下降。

摘要:

我们提出了一个用于评估深度搜索的新基准,深度搜索是检索增强生成(RAG)的一种现实且复杂的形式,需要在对多样、稀疏但相关的来源进行感知来源的多跳推理。这些来源包括文档、会议记录、Slack消息、GitHub和URL,它们在结构上各不相同,且通常包含人与人之间的交互。我们使用一个合成数据管道来构建它,该管道模拟了产品规划、开发和支持阶段的业务工作流,生成具有现实噪声的互联内容以及带有保证真实答案的多跳问题。我们发布的基准包含可回答和不可回答的查询,以及包含39,190个企业工件的检索池,从而能够对长上下文LLM和RAG系统进行细粒度评估。我们的实验表明,即使是性能最好的智能体RAG方法,在我们的基准上也仅达到32.96的平均性能得分。通过进一步分析,我们强调检索是主要瓶颈:现有方法难以进行深度搜索并检索所有必要的证据。因此,它们通常在部分上下文上进行推理,导致显著的性能下降。

英文摘要:

We present a new benchmark for evaluating Deep Search--a realistic and complex form of retrieval-augmented generation (RAG) that requires source-aware, multi-hop reasoning over diverse, sparsed, but related sources. These include documents, meeting transcripts, Slack messages, GitHub, and URLs, which vary in structure and often contain human-to-human interactions. We build it using a synthetic data pipeline that simulates business workflows across product planning, development, and support stages, generating interconnected content with realistic noise and multi-hop questions with guaranteed ground-truth answers. We release our benchmark with both answerable and unanswerable queries, and retrieval pool of 39,190 enterprise artifacts, enabling fine-grained evaluation of long-context LLM and RAG systems. Our experiments reveal that even the best-performing agentic RAG methods achieve an average performance score of 32.96 on our benchmark. With further analysis, we highlight retrieval as the main bottleneck: existing methods struggle to conduct deep searches and retrieve all necessary evidence. Consequently, they often reason over partial context, leading to significant performance degradation.

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。

↑