让AI理解文档布局和跨页关系，动态精准找证据。
LAD-RAG: Layout-aware Dynamic RAG for Visually-Rich Document Understanding
- 构建文档图谱保留版面结构与跨页依赖
- 动态检索证据，平均完美召回超90%
- 适合需要多页推理的文档问答任务
对视觉丰富文档（VRDs）的问答需同时推理内容与文档结构及跨页依赖。传统RAG在摄入时将内容切块独立编码，丢失结构与跨页关系，且推理时固定检索页面数，无法适配问题需求，导致证据不全、答案质量下降。为此，提出LAD-RAG：摄入阶段构建符号化文档图谱，融合神经嵌入，形成更完整的文档表征；推理阶段，由LLM代理动态交互神经与符号索引，根据查询自适应检索必要证据。在MMLongBench-Doc、LongDocURL、DUDE和MP-DocVQA上实验表明，LAD-RAG实现平均超过90%的完美召回率，无需top-k调优，相比基线在相近噪声水平下召回率提升最高达20%，显著提升问答准确率且延迟极低。
原文摘要 · Abstract (English)
Question answering over visually rich documents (VRDs) requires reasoning not only over isolated content but also over documents' structural organization and cross-page dependencies. However, conventional retrieval-augmented generation (RAG) methods encode content in isolated chunks during ingestion, losing structural and cross-page dependencies, and retrieve a fixed number of pages at inference, regardless of the specific demands of the question or context. This often results in incomplete evidence retrieval and degraded answer quality for multi-page reasoning tasks. To address these limitations, we propose LAD-RAG, a novel Layout-Aware Dynamic RAG framework. During ingestion, LAD-RAG constructs a symbolic document graph that captures layout structure and cross-page dependencies, adding it alongside standard neural embeddings to yield a more holistic representation of the document. During inference, an LLM agent dynamically interacts with the neural and symbolic indices to adaptively retrieve the necessary evidence based on the query. Experiments on MMLongBench-Doc, LongDocURL, DUDE, and MP-DocVQA demonstrate that LAD-RAG improves retrieval, achieving over 90% perfect recall on average without any top-k tuning, and outperforming baseline retrievers by up to 20% in recall at comparable noise levels, yielding higher QA accuracy with minimal latency.
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。