AnyDoc：通过大规模 HTML/CSS 数据合成与高度感知强化优化增强文档生成
AnyDoc: Enhancing Document Generation via Large-Scale HTML/CSS Data Synthesis and Height-Aware Reinforcement Optimization
提出可扩展的 HTML/CSS 文档数据合成管线与高度感知强化学习后训练，为多模态大语言模型在结构化视觉生成任务（如文档渲染、元素到文档）上提供可复用的数据构造与奖励设计思路。
原文依据 AnyDoc first establishes a scalable data synthesis pipeline to automatically generate documents in HTML/CSS form.
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生成在人工智能驱动的内容创作领域日益受到关注。本文通过引入 AnyDoc 框架推动其边界，该框架能够处理广泛文档类别中的多种生成任务，且所有文档均以统一的 HTML/CSS 格式表示。为克服现有手工构建文档数据集在覆盖范围和规模上的局限，AnyDoc 首先建立了一个可扩展的数据合成管线，以自动生成 HTML/CSS 形式的文档。该管线产出了 DocHTML，一个包含 265,206 个文档样本的大规模数据集，涵盖 111 个类别和 32 种不同风格。此外，所有文档均配备了全面的元数据，包括设计意图、HTML/CSS 源代码、视觉资产和渲染截图。基于该精选数据集，AnyDoc 微调多模态大语言模型（MLLMs）以实现三种实用的文档生成任务：意图到文档、文档反渲染和元素到文档。为解决微调过程中观察到的内容溢出问题，AnyDoc 进一步引入了高度感知强化学习（HARL）后训练流程。通过定义基于预测文档高度与目标文档高度差异的奖励函数，溢出在 HARL 过程中受到惩罚并逐渐缓解，从而提升整体性能。定性和定量实验表明，AnyDoc 在所有三个任务上均优于通用 MLLMs 和任务特定基线。
英文原文摘要 1547 字符
Document generation has gained growing attention in the field of AI-driven content creation. In this work, we push its boundaries by introducing AnyDoc, a framework capable of handling multiple generation tasks across a wide spectrum of document categories, all represented in a unified HTML/CSS format. To overcome the limited coverage and scale of existing human-crafted document datasets, AnyDoc first establishes a scalable data synthesis pipeline to automatically generate documents in HTML/CSS form. This pipeline yields DocHTML, a large-scale dataset containing 265,206 document samples, while spanning 111 categories and 32 distinct styles. Additionally, all documents are equipped with comprehensive metadata, including design intentions, HTML/CSS source code, visual assets, and rendered screenshots. Building on the curated dataset, AnyDoc fine-tunes multi-modal large language models (MLLMs) to achieve three practical document generation tasks: intention-to-document, document derendering, and element-to-document. To address the content overflow issue observed during fine-tuning, AnyDoc further incorporates a height-aware reinforcement learning (HARL) post-training procedure. By defining a reward function based on the difference between predicted and target document heights, overflow is penalized and gradually mitigated during HARL, thereby enhancing overall performance. Qualitative and quantitative experiments demonstrate that AnyDoc outperforms both general-purpose MLLMs and task-specific baselines across all three tasks.