arXiv cs.CL· Sangjune An, Jaeyoon Jung, Yumin Choi, Jinheon Baek, Junmo Cho, Sung Ju Hwang, Hae Beom Lee·· 13 小时前AI 评分38
SearchWeaver：让智能体自动优化搜索接口
Let Agents Design How They Search: Automatic Interface Optimization for Agentic Search
AI 导读
研究者提出 SearchWeaver 框架，用执行轨迹和任务奖励自动优化搜索接口，联合修订语料产物、可执行工具和使用说明。搭配 GPT-5.4-nano 时，优化后的接口在 BrowseComp-Plus 上把答案准确率从 71.6% 提升到 78.4%，在四个 BRIGHT 子集上平均 nDCG@10 从 45.7 升至 55.0，均优于 Direct Corpus Interaction。
正文 · 原文
Abstract:Agentic search depends on the interface through which a language-model agent accesses and explores a corpus. Existing interface designs are fixed by hand and applied unchanged to every corpus, although effective designs may depend on the corpus. We introduce SearchWeaver, a framework for automatic search interface optimization guided by downstream search performance. A proposer uses execution trajectories and task rewards to jointly revise corpus artifacts, executable tools, and usage instructions. This pro cess optimizes what information is prepared in advance, which operations tools perform, and which decisions remain with the search agent. With GPT-5.4-nano, the optimized interfaces improve answer accuracy from 71.6% to 78.4% on BrowseComp-Plus and average nDCG@10 from 45.7 to 55.0 across four BRIGHT subsets over Direct Corpus Interaction. These findings highlight the value of jointly optimizing corpus artifacts, executable tools, and usage instructions to develop effective search interfaces.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.29250 [cs.CL] |
| (or arXiv:2610.29250v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.29250 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Sangjune An [view email]
[v1]
Wed, 21 Oct 2026 05:56:30 UTC (1,004 KB)
来源：arXiv cs.CL · arxiv.org