arXiv cs.CL· Ayoub El Bouchtili, Guilhaume Leroy-Meline·· 12 小时前AI 评分29
FRAGMENT：基于实体感知变换的因子化图表示用于文档生成与编辑
FRAGMENT: Factorized Graph Representations for Document Generation and Editing via Entity-Aware Transformations
AI 导读
FRAGMENT 将文档表示为带类型的关系图，并把分布因子化为 p(structure)·p(content|structure)，由因果掩码 Transformer 的 Architect 生成图拓扑与空间关系、GATv2 的 Builder 补全归一化边界框、文本与视觉样式。
正文 · 原文
Abstract:Structured documents such as invoices, forms, reports, and scientific articles derive meaning from the interplay between spatial layout, textual content, and logical structure. Generative models operating at the pixel or token level often struggle to capture these dependencies effectively. We explore FRAGMENT, a generative framework that represents a document as a typed relational graph and factorizes its distribution as p(structure, content) = p(structure) * p(content | structure). The framework consists of two stages. The first stage, the Architect, is a causally masked Transformer conditioned on document category that autoregressively generates the graph topology and typed spatial relations. The second stage, the Builder, is a GATv2-based graph attention network that enriches the graph with normalized bounding boxes, text, and visual style attributes. Both stages define explicit likelihood models, yielding a tractable document-level likelihood that serves as an anomaly score for forgery detection. For controlled editing, a prompt-conditioned extension injects instruction embeddings into the Builder through cross-attention, enabling semantic and entity-aware modifications. We describe training on DocLayNet and fine-tuning on FUNSD and SROIE. Experiments on DocLayNet, FUNSD, and SROIE evaluate FRAGMENT alongside representative autoregressive, layout-only, and graph-based baselines, providing an empirical analysis of the characteristics and trade-offs of the proposed factorized graph generation framework.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL) |
| Cite as: | arXiv:2608.18679 [cs.CV] |
| (or arXiv:2608.18679v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2608.18679 arXiv-issued DOI via DataCite |
Submission history
From: Ayoub El Bouchtili [view email]
[v1]
Wed, 19 Aug 2026 08:29:19 UTC (2,231 KB)
来源：arXiv cs.CL · arxiv.org