ViDR: Grounding Multimodal Deep Research Reports in Source Visual Evidence

2026-05-13

AI Summary

This work addresses the limitation of existing deep research report systems that predominantly rely on textual evidence while overlooking the potential of source images as verifiable visual evidence. The authors propose ViDR, a novel framework that systematically integrates source images as core evidence—making them retrievable, interpretable, routable, and verifiable. ViDR constructs an indexed, multimodal outline by leveraging context-aware filtering, outline-aware reranking, and vision-language model analysis, enabling the generation of long-form reports grounded in chapter-level multimodal evidence. Evaluated on the newly introduced MMR Bench+ benchmark, ViDR significantly outperforms current commercial and open-source baselines in terms of report quality, image integration, and verifiability.

Abstract

Recent deep research systems have improved the ability of large language models to produce long, grounded reports through iterative retrieval and reasoning. However, most text-centered systems rely mainly on textual evidence, while multimodal systems often retrieve images only weakly or generate charts themselves, leaving source figures underused as evidence. We present ViDR, a multimodal deep research framework that grounds long-form reports in source figures. ViDR treats source figures as retrievable, interpretable, routable, and verifiable evidence objects, while still generating analytical charts when needed. It builds an evidence-indexed outline linking claims to textual and visual evidence, refines noisy web images into source-figure evidence atoms through context-aware filtering, outline-aware reranking, and VLM-based visual analysis, and generates each section with section-specific evidence. ViDR further validates visual references to reduce hallucinated or misplaced figures. We also introduce MMR Bench+, a benchmark for evaluating visual evidence use in deep research reports, covering source-figure retrieval, placement, interpretation, verifiability, and analytical chart generation. Experiments show that ViDR improves overall report quality, source-figure integration, and verifiability over strong commercial and open-source baselines. These results suggest that source visual evidence is important for multimodal deep research, as it strengthens evidential grounding, visual support, and report verifiability.

Problem

• multimodal deep research
• source visual evidence
• grounded report generation
• figure retrieval
• evidential grounding

Innovation

• multimodal deep research
• source visual evidence
• evidence grounding
• visual language model
• report verifiability