{"ID":2836022,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.22521","arxiv_id":"2511.22521","title":"DocVAL: Validated Chain-of-Thought Distillation for Grounded Document VQA","abstract":"Document visual question answering requires models not only to answer questions correctly, but also to precisely localize answers within complex document layouts. While large vision-language models (VLMs) achieve strong spatial grounding, their inference cost and latency limit real-world deployment. Compact VLMs are more efficient, but they often suffer substantial localization degradation under standard fine-tuning or distillation. To address this gap, we propose DocVAL, a validated chain-of-thought (CoT) distillation framework that transfers explicit spatial reasoning from large teacher models to compact, deployable student VLMs. DocVAL combines (1) teacher-generated spatial CoT supervision, (2) a rule-based dual-mode validator that filters low-quality training signals and provides fine-grained, pixel-level corrective feedback, and (3) a validation-driven two-stage training procedure with iterative refinement. Text detection is used only as training-time scaffolding for supervision and validation, enabling the final student to operate as a pure VLM without OCR or detection at inference. Across multiple document understanding benchmarks, DocVAL yields consistent improvements of up to 6-7 ANLS points over comparable compact VLMs. We further introduce mean Average Precision (mAP) as a localization metric for document question answering and report strong spatial grounding performance under this new evaluation. We release 95K validator-verified CoT traces and show that high-quality, validated supervision is more effective than scaling unfiltered data, enabling efficient and trustworthy document grounding. Code/Data: https://github.com/ahmad-shirazi/DocVAL","short_abstract":"Document visual question answering requires models not only to answer questions correctly, but also to precisely localize answers within complex document layouts. While large vision-language models (VLMs) achieve strong spatial grounding, their inference cost and latency limit real-world deployment. Compact VLMs are mo...","url_abs":"https://arxiv.org/abs/2511.22521","url_pdf":"https://arxiv.org/pdf/2511.22521v3","authors":"[\"Pinaki Prasad Guha Neogi\",\"Ahmad Mohammadshirazi\",\"Ser-Nam Lim\",\"Rajiv Ramnath\"]","published":"2025-11-27T15:00:58Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.AI\"]","methods":"[\"Language Model\"]","has_code":false,"code_links":[{"ID":606563,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2836022,"paper_url":"https://arxiv.org/abs/2511.22521","paper_title":"DocVAL: Validated Chain-of-Thought Distillation for Grounded Document VQA","repo_url":"https://github.com/ahmad-shirazi/DocVAL","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}