用双图结构提升多模态文档问答的准确率
$G^2$-Reader: Dual Evolving Graphs for Multimodal Document QA
- 构建内容图与规划图双系统，保持文档结构和跨模态语义
- 在五类多模态数据上平均准确率达66.21%，优于GPT-5的53.08%
- 适合需要精准理解图文混排长文档的研究者与工程师
检索增强生成是长文档问答的实用范式，但在文本、表格、图表交错的多模态阅读中仍显脆弱。首先，扁平化分块破坏文档原生结构和跨模态对齐，产生难以独立理解的语义片段；其次，即使迭代检索也会因噪声累积而循环于部分证据或漂移至无关段落，因每一步仅依赖当前片段而无持久全局搜索状态。我们提出$G^2$-Reader，一个双图系统：演化内容图以保留文档原生结构与跨模态语义，维护规划图——一种用于追踪中间发现并引导分步导航以完成证据获取的智能有向无环图。在涵盖五个多模态领域的VisDoMBench上，$G^2$-Reader结合Qwen3-VL-32B-Instruct达到66.21%平均准确率，超越强基线及独立GPT-5（53.08%）。
原文摘要 · Abstract (English)
Retrieval-augmented generation is a practical paradigm for question answering over long documents, but it remains brittle for multimodal reading where text, tables, and figures are interleaved across many pages. First, flat chunking breaks document-native structure and cross-modal alignment, yielding semantic fragments that are hard to interpret in isolation. Second, even iterative retrieval can fail in long contexts by looping on partial evidence or drifting into irrelevant sections as noise accumulates, since each step is guided only by the current snippet without a persistent global search state. We introduce $G^2$-Reader, a dual-graph system, to address both issues. It evolves a Content Graph to preserve document-native structure and cross-modal semantics, and maintains a Planning Graph, an agentic directed acyclic graph of sub-questions, to track intermediate findings and guide stepwise navigation for evidence completion. On VisDoMBench across five multimodal domains, $G^2$-Reader with Qwen3-VL-32B-Instruct reaches 66.21\% average accuracy, outperforming strong baselines and a standalone GPT-5 (53.08\%).
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。