---
title: "From Recognition to Reasoning — paper by Wang et al. (2026)"
description: "The authors introduce ReceiptBench, a human-annotated collection of 10,656 real-world financial documents, including taxi… 38 references."
authors: ["Yandi Wang", "Libin Zhan", "Ziwei Huang", "Tiancheng Luo", "Yuxuan Jiang", "Wang Dong", "Leilei Gan", "Jun Chen"]
year: 2026
venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/2026.acl-long.2135/"
doi: "10.18653/v1/2026.acl-long.2135"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/wang-2026-recognition-reasoning-benchmarking-enhancing-mllms/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/wang-2026-recognition-reasoning-benchmarking-enhancing-mllms/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/34/) / From Recognition to Reasoning: Benchmarking and Enhancing MLLMs on Real-World Receipt Document Understanding

# From Recognition to Reasoning: Benchmarking and Enhancing MLLMs on Real-World Receipt Document Understanding

**Authors:** [Yandi Wang](https://md.research.elamur.ai/free/en/authors/yandi-wang/), Libin Zhan, [Ziwei Huang](https://md.research.elamur.ai/free/en/authors/ziwei-huang/), Tiancheng Luo, [Yuxuan Jiang](https://md.research.elamur.ai/free/en/authors/yuxuan-jiang/), Wang Dong, [Leilei Gan](https://md.research.elamur.ai/free/en/authors/leilei-gan/) and [Jun Chen](https://md.research.elamur.ai/free/en/authors/jun-chen/) · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.acl-long.2135/) · [PDF](https://aclanthology.org/2026.acl-long.2135.pdf) · [DOI 10.18653/v1/2026.acl-long.2135](https://doi.org/10.18653/v1/2026.acl-long.2135) · [OpenAlex](https://openalex.org/W7166831607)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary

The paper addresses Visual Information Extraction (VIE) from receipts and invoices. Existing benchmarks are small, synthetic, or limited to narrow domains, and they mostly test literal text extraction. The authors introduce ReceiptBench, a human-annotated collection of 10,656 real-world financial documents, including taxi invoices, ferry tickets and hotel statements. The documents were gathered through web crawling and paid crowdsourcing and labeled with 19 fine-grained fields. Tasks are organized into four levels: Basic Perception, Format Normalization, Semantic Reasoning and Structure Parsing. Scoring goes beyond string matching, using LLM-based judges and Hungarian matching for nested line items.

To improve open-source MLLMs, the authors propose two training stages. Supervised fine-tuning comes first. It is followed by Metric-Aware GRPO, whose reward engine turns evaluation rules such as schema checks and arithmetic consistency into reinforcement learning signals. The reward explicitly penalizes hallucinated values. According to the authors, the approach brings Qwen3-VL close to proprietary systems and even exceeds them on reasoning-heavy fields. Nested structure parsing remains unsolved. The released benchmark offers a stricter testbed for automated financial document processing.

## Key points

- **ReceiptBench contents:** 10,656 images and 19 fields, with data and code released on GitHub. The dataset is 97.9% English, and personal information is masked.
- **Reward design:** the engine combines rule checkers, LLM judges and list matchers. Outcomes are mapped to a 2x2 matrix that rewards true positives and penalizes false positives.
- **Ablation results (Qwen3-VL-4B):** SFT + GRPO scores Perception 0.8226 and Normalization 0.9298, compared with a one-shot baseline of 0.6407 and 0.7503. GRPO alone gives the best Reasoning score (0.8560).
- **Hallucination reduction:** Metric-Aware GRPO cuts false positives in complex fields by up to 68.9%.
- **Error analysis:** models such as Gemini sometimes alter line-item values or invent tax entries so that the totals add up. Gemini also guesses currencies poorly, for example making 241 std_curr errors versus 58 for the fine-tuned Qwen.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Zhejiang University](https://md.research.elamur.ai/free/en/universities/zhejiang-university/) — University, Hangzhou, China · 224 works · 13,587 citations

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 3,486 works

🧭 Topics (1): [🧮 Reasoning](https://md.research.elamur.ai/free/en/topics/reasoning/) · 1,685 works

## Cite

Cite as: Yandi Wang, Libin Zhan, Ziwei Huang, Tiancheng Luo, Yuxuan Jiang, Wang Dong, Leilei Gan and Jun Chen. 2026. From Recognition to Reasoning: Benchmarking and Enhancing MLLMs on Real-World Receipt Document Understanding. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{wang2026recognition,
  title = {{From Recognition to Reasoning: Benchmarking and Enhancing MLLMs on Real-World Receipt Document Understanding}},
  author = {Yandi Wang, Libin Zhan, Ziwei Huang, Tiancheng Luo, Yuxuan Jiang, Wang Dong, Leilei Gan and Jun Chen},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.2135},
  url = {https://doi.org/10.18653/v1/2026.acl-long.2135}
}
```

- [← previous From Questions to Assessment Tuples: A Multi-Agent Framework with Bloom-Specialized Agents and Automated Verification](https://md.research.elamur.ai/free/en/papers/wong-2026-questions-assessment-tuples-multi-agent/)
- [next → From Regulatory Approvals to Patents: Cross-Domain Linking for Cardiovascular Device Traceability](https://md.research.elamur.ai/free/en/papers/yang-2026-regulatory-approvals-patents-cross-domain/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/34/)

## References

38 references, 18 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Abdelrahman Abdallah, Mohamed Mounis, Mahmoud Abdalla, Mahmoud SalahEldin Kasem, and 1 others. 2024. Receiptsense: Beyond traditional ocr -a dataset for receipt understanding. arXiv preprint arXiv:2406.04493. — [arXiv:2406.04493](https://arxiv.org/abs/2406.04493) [find](https://scholar.google.com/scholar?q=Abdelrahman+Abdallah%2C+Mohamed+Mounis%2C+Mahmoud+Abdalla%2C+Mahmoud+SalahEldin+Kasem%2C+and+1+others%2E+2024%2E+Receiptsense%3A+Beyond+traditional+ocr+%2Da+dataset+for+receipt+understanding%2E+arXiv+preprint+arXiv%3A240)
2. Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xiaodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, and 1 others. 2023. [Qwen technical report](https://md.research.elamur.ai/free/en/papers/bai-2023-qwen-technical-report/). arXiv preprint arXiv:2309.16609.
3. Shuai Bai, Yuxuan Cai, Ruizhe Chen, Keqin Chen, Xionghui Chen, Zesen Cheng, Lianghao Deng, Wei Ding, Chang Gao, Chunjiang Ge, and 1 others. 2025. [Qwen3-vl technical report](https://md.research.elamur.ai/free/en/papers/bai-2025-qwen3-vl-technical-report/). arXiv preprint arXiv:2511.21631.
4. Ali Furkan Biten, Ruben Tito, Andres Mafla, Lluis Gomez, Marçal Rusinol, Ernest Valveny, CV Jawahar, and Dimosthenis Karatzas. 2019. Scene text visual question answering. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 4291–4301. [find](https://scholar.google.com/scholar?q=Ali+Furkan+Biten%2C+Ruben+Tito%2C+Andres+Mafla%2C+Lluis+Gomez%2C+Mar%C3%A7al+Rusinol%2C+Ernest+Valveny%2C+CV+Jawahar%2C+and+Dimosthenis+Karatzas%2E+2019%2E+Scene+text+visual+question+answering%2E+In+Proceedings+of+the+IEEE%2FCV)
5. Lukas Blecher, Guillem Cucurull, Thomas Scialom, and Robert Stojnic. 2023. [Nougat: Neural optical understanding for academic documents](https://md.research.elamur.ai/free/en/papers/blecher-2023-nougat-neural-optical-understanding-academic/). arXiv preprint arXiv:2308.13418.
6. Qian Chen, Xianyin Zhang, Lifan Guo, Feng Chen, and Chi Zhang. 2025. Dianjin-ocr-r1: Enhancing ocr capabilities via a reasoning-and-tool interleaved vision-language model. arXiv preprint arXiv:2508.13238. — [arXiv:2508.13238](https://arxiv.org/abs/2508.13238) [find](https://scholar.google.com/scholar?q=Qian+Chen%2C+Xianyin+Zhang%2C+Lifan+Guo%2C+Feng+Chen%2C+and+Chi+Zhang%2E+2025%2E+Dianjin%2Docr%2Dr1%3A+Enhancing+ocr+capabilities+via+a+reasoning%2Dand%2Dtool+interleaved+vision%2Dlanguage+model%2E+arXiv+preprint+arXiv%3A2508%2E13)
7. Zhe Chen, Jiannan Wu, Wenhai Wang, Weijie Su, Guo Chen, Sen Xing, Muyan Zhong, Qinglong Zhang, Xizhou Zhu, Lewei Lu, and 1 others. 2024. [Internvl: Scaling up vision foundation models and aligning for generic visual-linguistic tasks](https://md.research.elamur.ai/free/en/papers/chen-2023-internvl-scaling-up-vision/). In Proceedings of the IEEE/CVF conference on computer vision and pa…
8. Council of the European Union. 2006. Council directive 2006/112/ec of 28 november 2006 on the common system of value added tax. [find](https://scholar.google.com/scholar?q=Council+of+the+European+Union%2E+2006%2E+Council+directive+2006%2F112%2Fec+of+28+november+2006+on+the+common+system+of+value+added+tax%2E)
9. Cheng Cui, Ting Sun, Suyin Liang, Tingquan Gao, Zelun Zhang, Jiaxuan Liu, Xueqing Wang, Changda Zhou, Hongen Liu, Manhui Lin, and 1 others. 2025. Paddleocr-vl: Boosting multilingual document parsing via a 0.9 b ultra-compact vision-language model. arXiv preprint arXiv:2510.14528. — [arXiv:2510.14528](https://arxiv.org/abs/2510.14528) [find](https://scholar.google.com/scholar?q=Cheng+Cui%2C+Ting+Sun%2C+Suyin+Liang%2C+Tingquan+Gao%2C+Zelun+Zhang%2C+Jiaxuan+Liu%2C+Xueqing+Wang%2C+Changda+Zhou%2C+Hongen+Liu%2C+Manhui+Lin%2C+and+1+others%2E+2025%2E+Paddleocr%2Dvl%3A+Boosting+multilingual+document+parsing+v)
10. Financial Accounting Standards Board (FASB). 2010. Statement of financial accounting concepts no. 8: Conceptual framework for financial reporting. Financial Accounting Series. Chapter 1: The Objective of General Purpose Financial Reporting. [find](https://scholar.google.com/scholar?q=Financial+Accounting+Standards+Board+%28FASB%29%2E+2010%2E+Statement+of+financial+accounting+concepts+no%2E+8%3A+Conceptual+framework+for+financial+reporting%2E+Financial+Accounting+Series%2E+Chapter+1%3A+The+Objective)
11. Anwen Hu, Haiyang Xu, Jiabo Ye, Ming Yan, Liang Zhang, Bo Zhang, Ji Zhang, Qin Jin, Fei Huang, and Jingren Zhou. 2024. [mplug-docowl 1.5: Unified structure learning for ocr-free document understanding](https://md.research.elamur.ai/free/en/papers/hu-2024-mplug-docowl-1-5-unified/). In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 3096–3120.
12. Mingxin Huang, Yongxin Shi, Dezhi Peng, Songxuan Lai, Zecheng Xie, and Lianwen Jin. 2025. Ocr-reasoning benchmark: Unveiling the true capabilities of mllms in complex text-rich image reasoning. arXiv preprint arXiv:2505.17163. — [arXiv:2505.17163](https://arxiv.org/abs/2505.17163) [find](https://scholar.google.com/scholar?q=Mingxin+Huang%2C+Yongxin+Shi%2C+Dezhi+Peng%2C+Songxuan+Lai%2C+Zecheng+Xie%2C+and+Lianwen+Jin%2E+2025%2E+Ocr%2Dreasoning+benchmark%3A+Unveiling+the+true+capabilities+of+mllms+in+complex+text%2Drich+image+reasoning%2E+arXiv+)
13. Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, and Furu Wei. 2022. [Layoutlmv3: Pre-training for document ai with unified text and image masking](https://md.research.elamur.ai/free/en/papers/huang-2022-layoutlmv3-pre-training-document-ai/). In Proceedings of the 30th ACM international conference on multimedia, pages 4083–4091.
14. Zheng Huang, Kai Chen, Jianhua He, Xiang Bai, Dimosthenis Karatzas, Shijian Lu, and CV Jawahar. 2019. Icdar2019 competition on scanned receipt ocr and information extraction. In Proceedings of the International Conference on Document Analysis and Recognition (ICDAR), pages 1516–1520. IEEE. [find](https://scholar.google.com/scholar?q=Zheng+Huang%2C+Kai+Chen%2C+Jianhua+He%2C+Xiang+Bai%2C+Dimosthenis+Karatzas%2C+Shijian+Lu%2C+and+CV+Jawahar%2E+2019%2E+Icdar2019+competition+on+scanned+receipt+ocr+and+information+extraction%2E+In+Proceedings+of+the+Int)
15. Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, and 1 others. 2024. [Gpt-4o system card](https://md.research.elamur.ai/free/en/papers/openai-2024-gpt-4o-system-card/). arXiv preprint arXiv:2410.21276.
16. Guillaume Jaume, Hazım Kemal Ekenel, and Jean-Philippe Thiran. 2019. Funsd: A dataset for form understanding in noisy scanned documents. In Proceedings of the International Conference on Document Analysis and Recognition Workshops (ICDARW), volume 2, pages 1–6. IEEE. [find](https://scholar.google.com/scholar?q=Guillaume+Jaume%2C+Haz%C4%B1m+Kemal+Ekenel%2C+and+Jean%2DPhilippe+Thiran%2E+2019%2E+Funsd%3A+A+dataset+for+form+understanding+in+noisy+scanned+documents%2E+In+Proceedings+of+the+International+Conference+on+Document+Anal)
17. Geewook Kim, Teakgyu Hong, Moonbin Yim, JeongYeon Nam, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, and Seunghyun Park. 2022. [Ocr-free document understanding transformer](https://md.research.elamur.ai/free/en/papers/kim-2021-ocr-free-document-understanding-transformer/). In European Conference on Computer Vision, pages 498–517. Springer.
18. Harold W Kuhn. 1955. The hungarian method for the assignment problem. Naval research logistics quarterly, 2(1-2):83–97. [find](https://scholar.google.com/scholar?q=Harold+W+Kuhn%2E+1955%2E+The+hungarian+method+for+the+assignment+problem%2E+Naval+research+logistics+quarterly%2C+2%281%2D2%29%3A83%E2%80%9397%2E)
19. Mahmoud Limam, Marwa Dhiaf, and Yousri Kessen-tini. 2025. Information extraction from multi-layout invoice images using fatura dataset. Engineering Applications of Artificial Intelligence, 149:110478. [find](https://scholar.google.com/scholar?q=Mahmoud+Limam%2C+Marwa+Dhiaf%2C+and+Yousri+Kessen%2Dtini%2E+2025%2E+Information+extraction+from+multi%2Dlayout+invoice+images+using+fatura+dataset%2E+Engineering+Applications+of+Artificial+Intelligence%2C+149%3A110478%2E)
20. Minesh Mathew, Dimosthenis Karatzas, and CV Jawahar. 2021. [Docvqa: A dataset for vqa on document images](https://md.research.elamur.ai/free/en/papers/mathew-2020-docvqa-dataset-vqa-document-images/). In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 2200–2209.
21. Seunghyun Park, Seung Shin, Bado Lee, Junyeop Lee, Jaeheung Surh, Minjoon Seo, and Hwalsuk Lee. 2019. Cord: A consolidated receipt dataset for post-ocr parsing. In Workshop on Document Intelligence at NeurIPS. [find](https://scholar.google.com/scholar?q=Seunghyun+Park%2C+Seung+Shin%2C+Bado+Lee%2C+Junyeop+Lee%2C+Jaeheung+Surh%2C+Minjoon+Seo%2C+and+Hwalsuk+Lee%2E+2019%2E+Cord%3A+A+consolidated+receipt+dataset+for+post%2Docr+parsing%2E+In+Workshop+on+Document+Intelligence+at)
22. Jake Poznanski, Aman Rangapur, Jon Borchardt, Jason Dunkelberger, Regan Huff, Daniel Lin, Christopher Wilhelm, Kyle Lo, and Luca Soldaini. 2025. [olmocr: Unlocking trillions of tokens in pdfs with vision language models](https://md.research.elamur.ai/free/en/papers/poznanski-2025-olmocr-unlocking-trillions-tokens-pdfs/). arXiv preprint arXiv:2502.18443.
23. Zhihong Shao, Peiyi Wang, Qihao Zhu, Runxin Xu, Junxiao Song, Xiao Bi, Haowei Zhang, Mingchuan Zhang, YK Li, Yang Wu, and 1 others. 2024. [Deepseekmath: Pushing the limits of mathematical reasoning in open language models](https://md.research.elamur.ai/free/en/papers/shao-2024-deepseekmath-pushing-limits-mathematical-reasoning/). arXiv preprint arXiv:2402.03300.
24. Hongbin Sun, Zhanghui Kuang, Xiaoyu Yue, Chenggang Lin, and Wayne Zhang. 2021. Spatial dual-modality graph reasoning for key information extraction. In Proceedings of the 29th ACM International Conference on Multimedia, pages 5229–5237. [find](https://scholar.google.com/scholar?q=Hongbin+Sun%2C+Zhanghui+Kuang%2C+Xiaoyu+Yue%2C+Chenggang+Lin%2C+and+Wayne+Zhang%2E+2021%2E+Spatial+dual%2Dmodality+graph+reasoning+for+key+information+extraction%2E+In+Proceedings+of+the+29th+ACM+International+Confer)
25. Rubèn Tito, Dimosthenis Karatzas, and Ernest Valveny. 2023. Hierarchical multimodal transformers for mul-tipage docvqa. Pattern Recognition, 144:109834. [find](https://scholar.google.com/scholar?q=Rub%C3%A8n+Tito%2C+Dimosthenis+Karatzas%2C+and+Ernest+Valveny%2E+2023%2E+Hierarchical+multimodal+transformers+for+mul%2Dtipage+docvqa%2E+Pattern+Recognition%2C+144%3A109834%2E)
26. Zhibo Yang, Jun Tang, Zhaohai Li, Pengfei Wang, Jianqiang Wan, Humen Zhong, Xuejing Liu, Mingkun Yang, Peng Wang, Shuai Bai, and 1 others. 2025. Cc-ocr: A comprehensive and challenging ocr benchmark for evaluating large multimodal models in literacy. In Proceedings of the IEEE/CVF International Conf… [find](https://scholar.google.com/scholar?q=Zhibo+Yang%2C+Jun+Tang%2C+Zhaohai+Li%2C+Pengfei+Wang%2C+Jianqiang+Wan%2C+Humen+Zhong%2C+Xuejing+Liu%2C+Mingkun+Yang%2C+Peng+Wang%2C+Shuai+Bai%2C+and+1+others%2E+2025%2E+Cc%2Docr%3A+A+comprehensive+and+challenging+ocr+benchmark+f)
27. Jiabo Ye, Anwen Hu, Haiyang Xu, Qinghao Ye, Ming Yan, Guohai Xu, Chenliang Li, Junfeng Tian, Qi Qian, Ji Zhang, and 1 others. 2023. [Ureader: Universal ocr-free visually-situated language understanding with multimodal large language model](https://md.research.elamur.ai/free/en/papers/ye-2023-ureader-universal-ocr-free-visually/). In Findings of the Association for Computational Linguistics:…
28. Jordy Van Landeghem, Rubèn Tito, Łukasz Borchmann, Michał Pietruszka, Pawel Joziak, Rafal Powalski, Dawid Jurkiewicz, Mickaël Coustaty, Bertrand Anck-aert, Ernest Valveny, and 1 others. 2023. Document understanding dataset and evaluation (dude). In Proceedings of the IEEE/CVF International Conferenc… [find](https://scholar.google.com/scholar?q=Jordy+Van+Landeghem%2C+Rub%C3%A8n+Tito%2C+%C5%81ukasz+Borchmann%2C+Micha%C5%82+Pietruszka%2C+Pawel+Joziak%2C+Rafal+Powalski%2C+Dawid+Jurkiewicz%2C+Micka%C3%ABl+Coustaty%2C+Bertrand+Anck%2Daert%2C+Ernest+Valveny%2C+and+1+others%2E+2023%2E+Document)
29. Yaowei Zheng, Richong Zhang, Junhao Zhang, Yanhan Ye, and Zheyan Luo. 2024. [Llamafactory: Unified efficient fine-tuning of 100+ language models](https://md.research.elamur.ai/free/en/papers/zheng-2024-llamafactory-unified-efficient-fine-tuning/). In Proceedings of the 62nd annual meeting of the association for computational linguistics (volume 3: system demonstrations), pages 400–410.
30. Jiapeng Wang, Chongyu Lian, Wenjing Wang, Xudong Ying, and Baoyuan Wang. 2021. Towards robust visual information extraction in real world: New dataset and novel solution. In Proceedings of the AAAI Conference on Artificial Intelligence. [find](https://scholar.google.com/scholar?q=Jiapeng+Wang%2C+Chongyu+Lian%2C+Wenjing+Wang%2C+Xudong+Ying%2C+and+Baoyuan+Wang%2E+2021%2E+Towards+robust+visual+information+extraction+in+real+world%3A+New+dataset+and+novel+solution%2E+In+Proceedings+of+the+AAAI+Co)
31. Xu Zhong, Jianbin Tang, and Antonio Jimeno Yepes. 2019. Publaynet: largest dataset ever for document layout analysis. In 2019 International Conference on Document Analysis and Recognition (ICDAR), pages 1015–1022. IEEE. [find](https://scholar.google.com/scholar?q=Xu+Zhong%2C+Jianbin+Tang%2C+and+Antonio+Jimeno+Yepes%2E+2019%2E+Publaynet%3A+largest+dataset+ever+for+document+layout+analysis%2E+In+2019+International+Conference+on+Document+Analysis+and+Recognition+%28ICDAR%29%2C+pag)
32. Haoran Wei, Chenglong Liu, Jinyue Chen, Jia Wang, Lingyu Kong, Yanming Xu, Zheng Ge, Liang Zhao, Jianjian Sun, Yuang Peng, and 1 others. 2024. General ocr theory: Towards ocr-2.0 via a unified end-to-end model. arXiv preprint arXiv:2409.01704. — [arXiv:2409.01704](https://arxiv.org/abs/2409.01704) [find](https://scholar.google.com/scholar?q=Haoran+Wei%2C+Chenglong+Liu%2C+Jinyue+Chen%2C+Jia+Wang%2C+Lingyu+Kong%2C+Yanming+Xu%2C+Zheng+Ge%2C+Liang+Zhao%2C+Jianjian+Sun%2C+Yuang+Peng%2C+and+1+others%2E+2024%2E+General+ocr+theory%3A+Towards+ocr%2D2%2E0+via+a+unified+end%2Dto%2D)
33. Jinguo Zhu, Weiyun Wang, Zhe Chen, Zhaoyang Liu, Shenglong Ye, Lixin Gu, Hao Tian, Yuchen Duan, Weijie Su, Jie Shao, and 1 others. 2025. [Internvl3: Exploring advanced training and test-time recipes for open-source multimodal models](https://md.research.elamur.ai/free/en/papers/zhu-2025-internvl3-exploring-advanced-training/). arXiv preprint arXiv:2504.10479.
34. Haoran Wei, Yaofeng Sun, and Yukun Li. 2025. [Deepseek-ocr: Contexts optical compression](https://md.research.elamur.ai/free/en/papers/wei-2025-deepseek-ocr-contexts-optical/). arXiv preprint arXiv:2510.18234.
35. Jason Wei, Maarten Bosma, Vincent Y Zhao, Kelvin Guu, Adams Wei Yu, Brian Lester, Nan Du, Andrew M Dai, and Quoc V Le. 2022a. [Finetuned language models are zero-shot learners](https://md.research.elamur.ai/free/en/papers/wei-2021-finetuned-language-models-zero-shot/). In International Conference on Learning Representations.
36. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, and 1 others. 2022b. [Chain-of-thought prompting elicits reasoning in large language models](https://md.research.elamur.ai/free/en/papers/wei-2022-chain-thought-prompting-elicits-reasoning/). Advances in neural information processing systems, 35:24824– 24837.
37. Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, and Ming Zhou. 2020. [Layoutlm: Pre-training of text and layout for document image understanding](https://md.research.elamur.ai/free/en/papers/xu-2019-layoutlm-pre-training-text-layout/). In Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pages 1192–1200.
38. Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, and Furu Wei. 2022. Xfund: A benchmark dataset for multilingual visually rich form understanding. In Findings of the Association for Computational Linguistics: ACL 2022, pages 3214–3224. [find](https://scholar.google.com/scholar?q=Yiheng+Xu%2C+Tengchao+Lv%2C+Lei+Cui%2C+Guoxin+Wang%2C+Yijuan+Lu%2C+Dinei+Florencio%2C+Cha+Zhang%2C+and+Furu+Wei%2E+2022%2E+Xfund%3A+A+benchmark+dataset+for+multilingual+visually+rich+form+understanding%2E+In+Findings+of+th)

*Updated 2 November 2026*