TVIR: Building Deep Research Agents Towards Text--Visual Interleaved Report Generation
2026-06-01 • Computation and Language
Computation and Language
AI summaryⓘ
The authors introduce TVIR, a new way to test and build AI systems that create detailed reports using both text and images. They created TVIR-Bench, a set of 100 tasks that need images to support the text analysis, and TVIR-Agent, a system that plans, finds images, makes charts, and writes reports carefully. They also designed a method to check both the text and visuals for accuracy. Their experiments show that TVIR-Agent performs well, highlighting the need for AI systems to handle both words and pictures when making evidence-based reports.
Deep Research AgentsMultimodal report generationImage retrievalChart generationHierarchical multi-agent systemsTextual assessmentVisual assessmentBenchmark datasetsContext-aware writingEvidence-driven analysis
Authors
Xinkai Ma, Zhiqi Bai, Dingling Zhang, Pei Liu, Yishuo Yuan, He Zhu, Jiakai Wang, Qianqian Xie, Yifan Zhao, Xinlong Yang, Hao Cong, Zhiheng Yao, Fengxia Xie, Zihao Xu, Haoran Xu, Zhaohui Wang, Minghao Liu, Shirong Lin, Yingshui Tan, Yuchi Xu, Wenbo Su, Zhaoxiang Zhang, Bo Zheng, Jiaheng Liu
Abstract
Deep Research Agents have shown strong capability in multi-step information retrieval, reasoning, and long-form report generation, but existing benchmarks and systems remain predominantly text-centric, with limited evaluation of whether visual elements are factually reliable and well aligned with the surrounding analysis. To address this gap, we introduce TVIR (Text--Visual Interleaved Report Generation), which includes TVIR-Bench, a benchmark of 100 expert-curated multimodal deep research tasks that require visual elements to serve specific analytical sub-goals, and TVIR-Agent, a hierarchical multi-agent framework that serves as a strong baseline for constructing outlines, retrieving images, generating charts with traceable sources, and composing reports through context-aware sequential writing. We further develop a dual-path evaluation framework that combines Textual Assessment and Visual Assessment. Experiments across nine deep research systems show that TVIR-Agent achieves strong overall performance, underscoring the importance of explicit multimodal design and evaluation for evidence-driven report generation.