HybridResearcher bridges evidence gap between structured and unstructured data
HybridResearcher: Bridging the Structured-Unstructured Boundary for Deep Research Agents
Artificial IntelligenceComputation and Language
Summary
Many questions in business or research need answers from both databases and web pages, but current AI agents struggle to combine these different sources well. The authors created HybridResearcher, a smaller AI model trained specifically to connect information across these kinds of data. They built training data that forces the AI to rely on both database queries and web searches together, then taught the AI with a mix of examples and reinforcement learning. This approach significantly improved its performance on a tough benchmark and also kept it good at working with just one type of data.
What this means in practice
- •For enterprise data teams: Combine internal database queries with web search results to answer complex business questions that span different data types.
- •For software developers: Build more capable AI assistants that can integrate structured and unstructured information without losing performance on single-source tasks.
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.