AI summaryⓘ
The authors introduce a new kind of deep research task called Knowledgeable Deep Research (KDR) that uses both structured data (like tables) and unstructured data (like text) to create detailed reports. They propose a system called Hybrid Knowledge Analysis (HKA) that uses multiple AI agents to analyze this mixed information, including text, images, and tables. To test their approach, they built a benchmark called KDR-Bench with complex questions and lots of data types. Their experiments show that HKA performs better than current systems, especially when handling visual information. The authors aim for this work to help future research that involves analyzing structured knowledge in AI-based deep research.
Deep ResearchLarge Language ModelsStructured KnowledgeUnstructured KnowledgeMultimodal ReportsHybrid Knowledge AnalysisBenchmarkVision-Language ModelsKnowledge-Centric Evaluation
Authors
Wenxuan Liu, Zixuan Li, Bai Long, Chunmao Zhang, Fenghui Zhang, Zhuo Chen, Wei Li, Yuxin Zuo, Fei Wang, Bingbing Xu, Xuhui Jiang, Jin Zhang, Xiaolong Jin, Jiafeng Guo, Tat-Seng Chua, Xueqi Cheng
Abstract
Deep Research (DR) requires LLM agents to autonomously perform multi-step information seeking, processing, and reasoning to generate comprehensive reports. In contrast to existing studies that mainly focus on unstructured web content, a more challenging DR task should additionally utilize structured knowledge to provide a solid data foundation, facilitate quantitative computation, and lead to in-depth analyses. In this paper, we refer to this novel task as Knowledgeable Deep Research (KDR), which requires DR agents to generate reports with both structured and unstructured knowledge. Furthermore, we propose the Hybrid Knowledge Analysis framework (HKA), a multi-agent architecture that reasons over both kinds of knowledge and integrates the texts, figures, and tables into coherent multimodal reports. The key design is the Structured Knowledge Analyzer, which utilizes both coding and vision-language models to produce figures, tables, and corresponding insights. To support systematic evaluation, we construct KDR-Bench, which covers 9 domains, includes 41 expert-level questions, and incorporates a large number of structured knowledge resources (e.g., 1,252 tables). We further annotate the main conclusions and key points for each question and propose three categories of evaluation metrics including general-purpose, knowledge-centric, and vision-enhanced ones. Experimental results demonstrate that HKA consistently outperforms most existing DR agents on general-purpose and knowledge-centric metrics, and even surpasses the Gemini DR agent on vision-enhanced metrics, highlighting its effectiveness in deep, structure-aware knowledge analysis. Finally, we hope this work can serve as a new foundation for structured knowledge analysis in DR agents and facilitate future multimodal DR studies.