M3Grounder：基于掩码的多跨度多粒度视觉定位用于文档问答
M3Grounder: Mask-Based Multi-Span and Multi-Granular Grounding for Document QA
将文档视觉定位从粗粒度边界框推进到像素级分割掩码，并引入多跨度、多粒度的层次化定位监督，为文档理解中的细粒度证据定位提供了可复用的分割式建模范式。
原文依据 M3Grounder, a hybrid vision-language and segmentation architecture that formulates document grounding as pixel-level segmentation.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论文档问答不仅需要准确的答案，还需要识别每个答案在页面上的定位位置。大多数模型将该任务视为纯文本生成，而现有的答案定位方法生成粗粒度的边界框，无法捕捉弯曲文本。我们提出 M3Grounder，一种混合视觉-语言与分割架构，将文档定位建模为像素级分割。它生成细粒度的证据掩码，并通过溢出抑制损失进行精炼以防止溢出。M3Grounder 自回归地生成答案文本，并与 [GROUND] 标记交错，将各个答案跨度链接到其对应的证据区域。此外，M3Grounder 使用强制空间包含的包围损失，在短语、行和块级别上层次化地定位证据。我们发布了 GroundingDocQA 数据集（20 万份文档、200 万多跨度多粒度问答对，带有像素级定位掩码），该数据集通过一个能够处理复杂布局、弯曲文本和图形丰富文档的数据引擎构建。我们还发布了 GroundingDocQA-Bench，一个多样且具有挑战性的人工验证基准。M3Grounder 在基于定位的文档视觉问答中树立了新的技术水平，从粗粒度边界框推进到层次化、细粒度且上下文关联的证据。
英文原文摘要 1308 字符
**Document QA** requires not only accurate answers but also identifying where each answer is grounded on the page. Most models treat the task as text-only generation, while existing answer grounding methods generate coarse bounding boxes that fail to capture curved text. We introduce **M3Grounder, a hybrid vision-language and segmentation architecture that formulates document grounding as pixel-level segmentation. It produces fine-grained evidence masks** refined by a bleed-suppression loss to prevent spillover. M3Grounder autoregressively generates answer text interleaved with [GROUND] tokens that link individual answer spans to their corresponding evidence regions. Also, **M3Grounder grounds evidence hierarchically across phrase, line, and block levels** using an enclosure loss that enforces spatial containment. We release **GroundingDocQA dataset (200K documents, 2M multi-span and multi-granular QA pairs with pixel-level grounding masks)**, built through a data engine that handles complex layouts, curved-text, and graphics-rich documents. We also release **GroundingDocQA-Bench, a diverse and challenging human-verified benchmark**. M3Grounder sets **a new state of the art in grounded DocVQA, advancing from coarse boxes to hierarchical, fine-grained and contextually grounded evidence.**