# DocAtlas: Multilingual Document Understanding Across 80+ Languages

arXiv: https://arxiv.org/abs/2605.12623 · HTML: https://mlsanity.com/paper/2605.12623/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 38 references that are in the map.
- First posted on arXiv 12 May 2026.

## Builds on (most cited first, 26 of 26)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency](https://mlsanity.com/paper/2508.18265/) (2025)
- [An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-tuning](https://mlsanity.com/paper/2308.08747/) (2023)
- [CCNet: Extracting High Quality Monolingual Datasets from Web Crawl Data](https://mlsanity.com/paper/1911.00359/) (2019)
- [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://mlsanity.com/paper/2204.08387/) (2022)
- [PubLayNet: largest dataset ever for document layout analysis](https://mlsanity.com/paper/1908.07836/) (2019)
- [FUNSD: A Dataset for Form Understanding in Noisy Scanned Documents](https://mlsanity.com/paper/1905.13538/) (2019)
- [FastText.zip: Compressing text classification models](https://mlsanity.com/paper/1612.03651/) (2016)
- [Nougat: Neural Optical Understanding for Academic Documents](https://mlsanity.com/paper/2308.13418/) (2023)
- [Image-based table recognition: data, model, and evaluation](https://mlsanity.com/paper/1911.10683/) (2019)
- [DocBank: A Benchmark Dataset for Document Layout Analysis](https://mlsanity.com/paper/2006.01038/) (2020)
- [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://mlsanity.com/paper/2104.08836/) (2021)
- [MonkeyOCR: Document Parsing with a Structure-Recognition-Relation Triplet Paradigm](https://mlsanity.com/paper/2506.05218/) (2025)
- [Dolphin: Document Image Parsing via Heterogeneous Anchor Prompting](https://mlsanity.com/paper/2505.14059/) (2025)
- [Docling: An Efficient Open-Source Toolkit for AI-driven Document Conversion](https://mlsanity.com/paper/2501.17887/) (2025)
- [LayoutParser: A Unified Toolkit for Deep Learning Based Document Image Analysis](https://mlsanity.com/paper/2103.15348/) (2021)
- [SmolDocling: An ultra-compact vision-language model for end-to-end multi-modal document conversion](https://mlsanity.com/paper/2503.11576/) (2025)
- [dots.ocr: Multilingual Document Layout Parsing in a Single Vision-Language Model](https://mlsanity.com/paper/2512.02498/) (2025)
- [SynthTIGER: Synthetic Text Image GEneratoR Towards Better Text Recognition Models](https://mlsanity.com/paper/2107.09313/) (2021)
- [PP-StructureV2: A Stronger Document Analysis System](https://mlsanity.com/paper/2210.05391/) (2022)
- [READoc: A Unified Benchmark for Realistic Document Structured Extraction](https://mlsanity.com/paper/2409.05137/) (2024)
- [KITAB-Bench: A Comprehensive Multi-Domain Benchmark for Arabic OCR and Document Understanding](https://mlsanity.com/paper/2502.14949/) (2025)
- [How to Teach Large Multimodal Models New Skills](https://mlsanity.com/paper/2510.08564/) (2025)
- [WordScape: a Pipeline to extract multilingual, visually rich Documents with Layout Annotations from Web Crawl Data](https://mlsanity.com/paper/2312.10188/) (2023)
- [Finetuning Vision-Language Models as OCR Systems for Low-Resource Languages: A Case Study of Manchu](https://mlsanity.com/paper/2507.06761/) (2025)

## Built on by (most cited first, 1 of 1)

- [Thinking with Anchors: Grounded and Efficient Document Reasoning](https://mlsanity.com/paper/2608.04424/) (2026)

## Abstract (arXiv)

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. Code is available at https://github.com/ahmedheakl/DocAtlas .