[
    {
        "id": "osp-15157",
        "type": "article-journal",
        "title": "SP-DocReader: Difference-Aware Self-Play for Precise Document OCR",
        "author": [
            {
                "family": "Liao",
                "given": "Wenjie"
            },
            {
                "family": "Song",
                "given": "Xiaohui"
            },
            {
                "family": "Zhao",
                "given": "Liangjie"
            },
            {
                "family": "Lu",
                "given": "Haonan"
            }
        ],
        "URL": "https://omanscience.com/en/articles/sp-docreader-difference-aware-self-play-for-precise-document-ocr",
        "language": "en",
        "issued": {
            "date-parts": [
                [
                    2026
                ]
            ]
        },
        "abstract": "Accurate page transcription remains difficult for vision language models under limited input and training budgets. We present SP-DocReader, a self-play framework for optical character recognition (OCR) that targets residual errors after supervised fine-tuning. Reading Discrepancy Masking aligns reference and generated model tokens through a longest common subsequence, then scores unmatched positions with their full conditioning prefixes. Focused Fidelity Loss adds direct negative log-likelihood supervision at unmatched ground-truth positions. Only the OCR module is trained, while the backbone remains frozen. We derive the combined gradient to distinguish relative score optimization from direct supervision. Compared with SFT-2, SP-DR-3 reduces Vary-600K character error rate on both backbones. On Qwen3-VL-4B, it reduces character error rate by approximately 54 percent and improves DocVQA Average Normalized Levenshtein Similarity (ANLS) by 3.7 points. These results show the value of focusing self-play training on the discrepancies that remain after supervised fine-tuning."
    }
]