@inproceedings{wang-etal-2026-infinity,
    title = "Infinity-Parser: Layout-Aware Reinforcement Learning with High-quality Document Parsing Dataset",
    author = "Wang, Baode  and
      Wu, Biao  and
      Li, Weizhen  and
      Fang, Meng  and
      Huang, Zuming  and
      Huang, Jun  and
      Liang, Yanjie  and
      Wang, Haozhe  and
      Chen, Ling  and
      Chu, Wei  and
      Qi, Yuan",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.82/",
    pages = "1647--1667",
    ISBN = "979-8-89176-395-1",
    abstract = "Document parsing from scanned images into structured formats remains a significant challenge due to its complexly intertwined elements such as text paragraphs, figures, formulas, and tables. Existing supervised fine-tuning methods often struggle to generalize across diverse document types, leading to poor performance, particularly on out-of-distribution data. This issue is further exacerbated by the limited availability of high-quality training data for layout-aware parsing tasks. To address these challenges, we introduce layoutRL, a reinforcement learning framework that optimizes layout understanding through composite rewards integrating normalized edit distance, paragraph count accuracy, and reading order preservation. To support this training, we construct the Infinity-Doc-400K dataset, which we use to train Infinity-Parser, a vision-language model demonstrating robust generalization across various domains. Extensive evaluations on benchmarks including OmniDocBench, olmOCR-Bench, PubTabNet, and FinTabNet show that Infinity-Parser consistently achieves state-of-the-art performance across a broad range of document types, languages, and structural complexities, substantially outperforming both specialized document parsing systems and general-purpose vision-language models. We will release our code, dataset, and model to facilitate reproducible research in document parsing."
}