看见即改进：用于迭代文本布局精修的视觉反馈
Seeing is Improving: Visual Feedback for Iterative Text Layout Refinement
把渲染后的图像作为反馈信号，通过带 OCR 准确率奖励的强化学习驱动 MLLM 迭代自我修正，这一「渲染-评估-重生成」闭环可直接迁移到任何需要视觉结果校验的生成式设计任务（如海报排版、UI 生成、图文合成）。
原文依据 we identify visual feedback as a critical factor in layout generation and propose Visual Feedback Layout Model (VFLM)
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论多模态大语言模型（MLLM）的最新进展使得从自然语言描述自动生成结构化布局成为可能。现有方法通常遵循仅代码范式，即生成代码来表示布局，再由图形引擎渲染以产生最终图像。然而，这些方法对渲染后的视觉结果不可见，因而难以保证可读性与美观性。本文指出视觉反馈是布局生成中的关键因素，并提出视觉反馈布局模型（VFLM），一个利用视觉反馈进行迭代精修的自改进框架。VFLM 能够执行自适应反思式生成，即利用视觉信息反思先前的问题，并迭代生成输出直至达到令人满意的质量。该能力通过强化学习实现，其奖励模型基于视觉并纳入 OCR 准确率。由于仅对最终生成结果给予奖励，我们能够有效激发模型的迭代与反思式生成能力。在多个基准上的实验表明，VFLM 持续优于先进的多模态大语言模型、现有布局模型以及仅代码基线，确立了视觉反馈对于面向设计的 MLLM 的关键作用。我们的代码与数据见 https://github.com/FolSpark/VFLM。
英文原文摘要 1383 字符
Recent advances in Multimodal Large Language Models (MLLMs) have enabled automated generation of structured layouts from natural language descriptions. Existing methods typically follow a code-only paradigm that generates code to represent layouts, which are then rendered by graphic engines to produce final images. However, they are blind to the rendered visual outcome, making it difficult to guarantee readability and aesthetics. In this paper, we identify visual feedback as a critical factor in layout generation and propose Visual Feedback Layout Model (VFLM), a self-improving framework that leverages visual feedback iterative refinement. VFLM is capable of performing adaptive reflective generation, which leverages visual information to reflect on previous issues and iteratively generates outputs until satisfactory quality is achieved. It is achieved through reinforcement learning with a visually grounded reward model that incorporates OCR accuracy. By rewarding only the final generated outcome, we can effectively stimulate the model's iterative and reflective generative capabilities. Experiments across multiple benchmarks show that VFLM consistently outperforms advanced MLLMs, existing layout models, and code-only baselines, establishing visual feedback as critical for design-oriented MLLMs. Our code and data are available at https://github.com/FolSpark/VFLM.