研究人员开发了 FLOWREADER，一种用于长篇多模态文档问答的新颖方法。该方法将证据组装重构为最小费用流问题，能够更好地处理跨文本、表格和幻灯片中的碎片化信息。在 VisDoMBench 基准测试的特定子集上，FLOWREADER 的表现优于传统的 top-k 检索方法，证明了其在复杂证据组装场景中的有效性。 AI
影响 引入了一种新颖的多模态问答方法，有望提高在复杂文档上的性能。
排序理由 该集群包含一篇详细介绍新的问答方法的学术论文。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2606.07235v1 Announce Type: cross Abstract: Long, multimodal documents force retrieval-augmented systems to assemble answers from evidence fragmented across text, tables, and slides broken across cells in a long table, spread over multiple slides, or split between a figure …
Long, multimodal documents force retrieval-augmented systems to assemble answers from evidence fragmented across text, tables, and slides broken across cells in a long table, spread over multiple slides, or split between a figure and its discussion. Top-$k$ chunk retrieval treats…
Long, multimodal documents force retrieval-augmented systems to assemble answers from evidence fragmented across text, tables, and slides broken across cells in a long table, spread over multiple slides, or split between a figure and its discussion. Top-$k$ chunk retrieval treats…