| title | DocAtlas: Multilingual Document Understanding Across 80+ Languages |
|---|---|
| authors | Ahmed Heakl, Youssef Mohamed, Abdullah Sohail, Rania Elbadry, Ahmed Nassar, Peter W. J. Staar, Fahad Shahbaz Khan, Imran Razzak, Salman Khan |
| date | 2026-05-12 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2605.12623 |
| affiliation | Mohamed Bin Zayed University of Artificial Intelligence |
| upvotes | 4 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | Multilingual document understanding remains limited for low-resource languages due to scarce training data and model-based annotation pipelines that perpetuate existing biases. We introduce DocAtlas, a framework that constructs high-fidelity OCR datasets and benchmarks covering 82 languages and 9 evaluation tasks. Our dual pipelines, differential rendering of native DOCX documents and synthetic LaTeX-based generation for right-to-left scripts produce precise structural annotations in a unified DocTag format encoding layout, text, and component types, without learned models for core annotation. Evaluating 16 state-of-the-art models reveals persistent gaps in low-resource scripts. We show that Direct Preference Optimization (DPO) using rendering-derived ground truth as positive signal achieves stable multilingual adaptation, improving both in-domain (+1.9%) and out-of-domain (+1.8%) accuracy without measurable base-language degradation, where supervised fine-tuning degrades out-of-domain performance by up to 21%. Our best variant, DocAtlas-DeepSeek, improves +1.7% over the strongest baseline. |
cite
@misc{heakl2026docatlas,
title={DocAtlas: Multilingual Document Understanding Across 80+ Languages},
author={Ahmed Heakl and Youssef Mohamed and Abdullah Sohail and Rania Elbadry and Ahmed Nassar and Peter W. J. Staar and Fahad Shahbaz Khan and Imran Razzak and Salman Khan},
year={2026},
eprint={2605.12623},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2605.12623}
}