arXiv cs.CL· Soyoung Yoon, Peiran Xu, Han Wang, Ruofan Wu, Xiaodong Yu, Feng Yan, Mohit Bansal, Debiao Li, Seung-Won Hwang, Yuxiong He, Zhewei Yao·· 13 小时前AI 评分34
HybridResearcher：让 4B 智能体打通结构化与非结构化数据边界
HybridResearcher: Bridging the Structured-Unstructured Boundary for Deep Research Agents
AI 导读
研究团队提出 HybridResearcher，一个 4B 参数智能体，专门训练跨结构化与非结构化数据源的证据整合能力。在 HybridDeepResearch 基准上，其 pass@8 比基座模型 Qwen3.5-4B 提升 30.0 分，并超过更大的 Qwen3.5-397B-A17B 达 8.3 分。
正文 · 原文
Authors:Soyoung Yoon, Peiran Xu, Han Wang, Ruofan Wu, Xiaodong Yu, Feng Yan, Mohit Bansal, Debiao Li, Seung-Won Hwang, Yuxiong He, Zhewei Yao
Abstract:While LLM agents are increasingly capable of gathering and synthesizing evidence from a single homogeneous source, such as structured tables or unstructured data, real-world analytical questions frequently span heterogeneous sources. For example, answering one enterprise question may require both a result computed over an internal database and evidence found on the web. Recent evaluations on benchmarks like HybridDeepResearch show that even frontier agents struggle when evidence must cross this boundary. While these benchmarks have exposed the limitation, methodologies to explicitly teach cross-source reasoning remain absent. To address this, we introduce HybridResearcher, a 4B agent trained to bridge evidence across the structured and unstructured boundary. To build it, we synthesize cross-source bridge trajectories by starting with validated single-source tasks-such as existing text-to-SQL or web search queries-and fabricating the missing source around a bridge entity drawn from the seed's own output. We retain a sample only when strict construction gates confirm that neither side alone suffices, ensuring the data requires true cross-source interdependence. Finally, we train on this curated dataset using supervised fine-tuning followed by reinforcement learning. Evaluated on the existing HybridDeepResearch benchmark, HybridResearcher improves over our base model, Qwen3.5-4B, by 30.0 points pass@8 and surpasses the far larger Qwen3.5-397B-A17B by 8.3 points, while also improving over its base model on established, held-out single-source benchmarks (BrowseComp and BrowseComp-Plus), demonstrating that this complex cross-source reasoning can be taught directly to small models without triggering catastrophic forgetting.
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.24899 [cs.AI] |
| (or arXiv:2610.24899v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.24899 arXiv-issued DOI via DataCite |
Submission history
From: Xiaodong Yu [view email]
[v1]
Sat, 17 Oct 2026 22:46:56 UTC (339 KB)
[v2]
Wed, 21 Oct 2026 00:41:07 UTC (324 KB)
来源：arXiv cs.CL · arxiv.org