TRivia：面向表格识别的视觉语言模型自监督微调
TRivia: Self-supervised Fine-tuning of Vision-Language Models for Table Recognition
把 GRPO 式的强化学习自监督微调用在 VLM 上，用「识别结果能否被正确问答」构造无需标注的奖励信号，并配注意力引导的多样化问题生成；这套闭环奖励设计可迁移到任何缺乏标注的结构化图像解析任务（如公式、图表、版面解析）。
原文依据 Built upon Group Relative Policy Optimization, TRivia automatically identifies unlabeled samples that most effectively facilitate learning
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论表格识别（TR）旨在将表格图像转换为 HTML 或 Markdown 等半结构化表示。作为文档解析的核心组成部分，表格识别长期依赖监督学习，近期的工作则以使用标注数据微调视觉语言模型（VLM）为主。尽管 VLM 将表格识别提升到了新的水平，但要进一步提升性能，需要大规模标注数据，而这类数据的获取成本高昂。因此，尽管闭源模型不断刷新性能边界，开源模型——通常以有限资源训练，且在实践中因隐私法规而成为许多用户唯一可行的选择——仍然远远落后。为弥合这一差距，我们提出 TRivia，一种自监督微调方法，使预训练的 VLM 能够直接从野外无标注表格图像中学习表格识别。TRivia 建立在组相对策略优化之上，能够自动识别最有助于学习的无标注样本，并通过基于问答的奖励机制免除对人工标注的需求。一个注意力引导模块为每张表格图像生成多样化的问题，而模型解读识别结果并正确回答这些问题的能力则提供反馈，用于优化表格识别模型。这一闭环过程使表格识别模型能够在无标注数据的情况下自主学习识别、结构化并推理表格。借助该流程，我们提出 TRivia-3B，一个开源、紧凑且达到最优水平的表格识别模型，在三个常用基准上超越了现有系统（如 Gemini 2.5 Pro、MinerU2.5）。
英文原文摘要 1616 字符
Table recognition (TR) aims to transform table images into semi-structured representations such as HTML or Markdown.As a core component of document parsing, TR has long relied on supervised learning, with recent efforts dominated by fine-tuning vision-language models (VLMs) using labeled data.While VLMs have brought TR to the next level, pushing performance further demands large-scale labeled data that is costly to obtain.Consequently, although proprietary models have continuously pushed the performance boundary, open-source models, often trained with limited resources and, in practice, the only viable option for many due to privacy regulations, still lag far behind.To bridge this gap, we introduce TRivia, a self-supervised fine-tuning method that enables pretrained VLMs to learn TR directly from unlabeled table images in the wild. Built upon Group Relative Policy Optimization, TRivia automatically identifies unlabeled samples that most effectively facilitate learning and eliminates the need for human annotations through a question-answering-based reward mechanism. An attention-guided module generates diverse questions for each table image, and the ability to interpret the recognition results and answer them correctly provides feedback to optimize the TR model.This closed-loop process allows the TR model to autonomously learn to recognize, structure, and reason over tables without labeled data. Leveraging this pipeline, we present TRivia-3B, an open-sourced, compact, and state-of-the-art TR model that surpasses existing systems (e.g., Gemini 2.5 Pro, MinerU2.5) on three popular benchmarks.