Elamur Research⌕
☰
home / papers / 2026 / Corpora Generation for Urdu Grammatical Error Correction

Corpora Generation for Urdu Grammatical Error Correction

Summary

Urdu has many millions of speakers but almost no annotated parallel data. As a result, automatic grammatical error correction (GEC) for this language has received little study, and neural seq2seq models lack training material. The authors propose a pipeline that synthesizes training pairs without manual annotation and without rules written specifically for Urdu. First, "before and after" sentence pairs are selected from the edit history of the Urdu Wikipedia. These pairs are analyzed with Stanza and a custom alignment algorithm inspired by ERRANT. Each error is stored as a morphosyntactic "core", meaning the changed word together with its nearest neighbors. These patterns are then transferred to grammatically correct sentences, which are artificially corrupted, and error frequencies are adjusted to match the natural distribution. This produced the UrduGEC-Synthetic corpus. Separately, a gold-standard test set, UrduGEC-Gold, was compiled from students' examination papers. Models such as mT0 and NLLB perform noticeably better on the Gold set after fine-tuning on the synthetic data. The authors believe the method can be transferred to other Indo-Aryan languages that have POS taggers and treebanks.

Key points

  • Wiki-Edits Dataset: about 240K sentence pairs from the edit history of the Urdu Wikipedia. The latest revision is treated as correct and the one before it as erroneous. After cleaning, the corpus retains only grammatical edits.
  • Error annotation: Stanza identifies POS tags, lemmas and morphological features. A closed vocabulary built from the UD 2.12 treebank checks words for OOV status so that purely orthographic corrections can be discarded. The alignment divides edits into I/D/S (insertion, deletion, substitution). Instead of the standard Levenshtein distance, it uses a linguistically motivated cost function.
  • Resources: UrduGEC-Synthetic contains 1.27 million sentence pairs. UrduGEC-Gold consists of digitized, anonymized student papers and is small. Both datasets are released on Hugging Face, and the pipeline code is on GitHub.
  • Result: after fine-tuning on UrduGEC-Synthetic, the mT0 and NLLB models substantially improved their results on UrduGEC-Gold.
  • Limitations: the core covers a three-word context, so the approach does not capture non-local or structural errors. Urdu has no counterpart to ERRANT, so a direct comparison with methods such as Choe et al. (2019) is not possible.

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

Fields: Linguistics & languages

Topics: Treebanks · Part-of-speech tagging

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.2156

@inproceedings{ahad2026corpora,
  title = {{Corpora Generation for Urdu Grammatical Error Correction}},
  author = {Ahad, Syed and Ezzi, Burhanuddin Aliasghar and Hussain, Muhammad Arsalan and Kumar, Sandesh and Samad, Abdul},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.2156},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.2156}
}

References

33 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Christopher Bryant, Zheng Yuan, Muhammad Reza Qorib, Hannan Cao, Hwee Tou Ng, and Ted Briscoe. 2023. Grammatical error correction: A survey of the state of the art. Computational Linguistics, 49(3):643–701. find
  2. Miriam Butt and Tafseer Ahmed. 2011. The redevelopment of indo-aryan case systems from a lexical semantic perspective. Morphology, 21(3):545–572. find
  3. Yo Joong Choe, Jiyeon Ham, Kyubyong Park, and Yeoil Yoon. 2019. A neural grammatical error correction system built on better pre-training and sequential transfer learning. In Proceedings of the Fourteenth Workshop on Innovative Use of NLP for Building Educational Applications, pages 213–227, Florenc… find
  4. Daniel Dahlmeier and Hwee Tou Ng. 2012. Better evaluation for grammatical error correction. In Proceedings of the 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 568–572, Montréal, Canada. Association for Computationa… find
  5. Mariano Felice and Zheng Yuan. 2014. Generating artificial errors for grammatical error correction. In Proceedings of the Student Research Workshop at the 14th Conference of the European Chapter of the Association for Computational Linguistics, pages 116–126, Gothenburg, Sweden. Association for Comp… find
  6. Jennifer Foster and Oistein Andersen. 2009. Gen-ERRate: Generating errors for use in grammatical error detection. In Proceedings of the Fourth Workshop on Innovative Use of NLP for Building Educational Applications, pages 82–90, Boulder, Colorado. Association for Computational Linguistics. find
  7. Google DeepMind. 2025. Gemini 3 Flash Model Card. Technical report, Google DeepMind. find
  8. Ahlam Alrehili and Areej Alhothali. 2025. Towards the development of balanced synthetic data for correcting grammatical errors in arabic: An approach based on error tagging model and synthetic data generating model. Preprint, arXiv:2502.05312. find
  9. Roman Grundkiewicz, Marcin Junczys-Dowmunt, and Kenneth Heafield. 2019. Neural grammatical error correction systems with unsupervised pre-training on synthetic data. In Proceedings of the Fourteenth Workshop on Innovative Use of NLP for Building Educational Applications, pages 252–263, Florence, Ita… find
  10. Riyaz Ahmad Bhat, Rajesh Bhatt, Annahita Farudi, Prescott Klassen, Bhuvana Narasimhan, Martha Palmer, Owen Rambow, Dipti Misra Sharma, Ashwini Vaidya, Sri Ramagurumurthy Vishnu, and Fei Xia. 2017. The hindi/urdu treebank project. In Nancy Ide and James Pustejovsky, editors, Handbook of Linguistic An… find
  11. Phu Mon Htut and Joel Tetreault. 2019. The unbearable weight of generating artificial errors for grammatical error correction. In Proceedings of the Fourteenth Workshop on Innovative Use of NLP for Building Educational Applications, pages 478–483, Florence, Italy. Association for Computational Lingu… find
  12. Christopher Bryant, Mariano Felice, and Ted Briscoe. 2017. Automatic annotation and evaluation of error types for grammatical error correction. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 793–805, Vancouver, Canada. Associ… find our card
  13. Marcin Junczys-Dowmunt, Roman Grundkiewicz, Shubha Guha, and Kenneth Heafield. 2018. Approaching neural grammatical error correction as a low-resource machine translation task. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human… find
  14. Sudhanshu Kasewa, Pontus Stenetorp, and Sebastian Riedel. 2018. Wronging a right: Generating better errors to improve grammatical error detection. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 4977–4983, Brussels, Belgium. Association for Computatio… find
  15. Masamune Kobayashi, Masato Mita, and Mamoru Komachi. 2024. Large language models are state-of-the-art evaluator for grammatical error correction. In Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024), pages 68–77, Mexico City, Mexico. Associati… find
  16. Philipp Koehn. 2004. Statistical significance tests for machine translation evaluation. In Proceedings of the 2004 Conference on Empirical Methods in Natural Language Processing, pages 388–395, Barcelona, Spain. Association for Computational Linguistics. find
  17. Niklas Muennighoff, Thomas Wang, Lintang Sutawika, Adam Roberts, Stella Biderman, Teven Le Scao, M Saiful Bari, Sheng Shen, Zheng-Xin Yong, Hailey Schoelkopf, et al. 2022. Crosslingual generalization through multitask finetuning. arXiv preprint arXiv:2211.01786. find our card
  18. Courtney Napoles, Keisuke Sakaguchi, Matt Post, and Joel Tetreault. 2016. Gleu without tuning. Preprint, arXiv:1605.02592. find
  19. Martha Palmer, Rajesh Bhatt, Bhuvana Narasimhan, Owen Rambow, Dipti Misra Sharma, and Fei Xia. 2009. Hindi syntax: Annotating dependency, lexical predicate-argument structure, and phrase structure. In The 7th International Conference on Natural Language Processing, pages 14–17. find
  20. Peng Qi, Yuhao Zhang, Yuhui Zhang, Jason Bolton, and Christopher D. Manning. 2020. Stanza: A Python natural language processing toolkit for many human languages. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: System Demonstrations. find our card
  21. Aomi Koyama, Kengo Hotate, Masahiro Kaneko, and Mamoru Komachi. 2021. Comparison of grammatical error correction using back-translation models. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Student Research Workshop, pages 126–1… find
  22. Marek Rei, Mariano Felice, Zheng Yuan, and Ted Briscoe. 2017. Artificial error generation with machine translation and syntactic patterns. In Proceedings of the 12th Workshop on Innovative Use of NLP for Building Educational Applications, pages 287–292, Copenhagen, Denmark. Association for Computati… find our card
  23. Jared Lichtarge, Chris Alberti, Shankar Kumar, Noam Shazeer, Niki Parmar, and Simon Tong. 2019. Corpora generation for grammatical error correction. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volu… find
  24. Grigori Sidorov, Anubhav Gupta, Martin Tozer, Dolors Catala, Angels Catena, and Sandrine Fuentes. 2013. Rule-based system for automatic grammar correction using syntactic n-grams for English language learning (L2). In Proceedings of the Seventeenth Conference on Computational Natural Language Learni… find
  25. Agnes Luhtaru, Taido Purason, Martin Vainikko, Maksym Del, and Mark Fishel. 2024. To err is human, but llamas can learn it too. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 12466–12481, Miami, Florida, USA. Association for Computational Linguistics. find
  26. Ankur Sonawane, Sujeet Kumar Vishwakarma, Bha-vana Srivastava, and Anil Kumar Singh. 2020. Generating inflectional errors for grammatical error correction in Hindi. In Proceedings of the 1st Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 10th Internat… find
  27. Shirong Ma, Yinghui Li, Rongyi Sun, Qingyu Zhou, Shulin Huang, Ding Zhang, Li Yangning, Ruiyang Liu, Zhongli Li, Yunbo Cao, Haitao Zheng, and Ying Shen. 2022. Linguistic rules-based corpus generation for native Chinese grammatical error correction. In Findings of the Association for Computational Li… find our card
  28. Riaz Ahmed Mangrio. 2016. The Morphology of Loanwords in Urdu: The Persian, Arabic and English Strands. Cambridge Scholars Publishing, Newcastle upon Tyne. find
  29. Riaz Ahmed Mangrio, Muhammad Javed Iqbal, Zafeer Hussain Kiani, and Rashida Imran. 2021. Morpho-phonological similarities in indo aryan languages: A descriptive account. Kashmir Journal of Language Research, 23(2). find
  30. NLLB Team, Marta R. Costa-jussà, James Cross, Onur Çelebi, Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean Maillard, Anna Sun, Skyler Wang, Guillaume Wenzek, Al Youngblood, Bapi Akula, Loic Barrault, Gabriel Mejia Gonzalez, Prangthip Hansanti, John Hof… find
  31. Ziang Xie, Guillaume Genthial, Stanley Xie, Andrew Ng, and Dan Jurafsky. 2018. Noising and denoising natural language: Diverse backtranslation for grammar correction. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language… find
  32. Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, and Colin Raffel. 2022. ByT5: Towards a token-free future with pre-trained byte-to-byte models. Transactions of the Association for Computational Linguistics, 10:291–306. find
  33. Zheng Yuan and Ted Briscoe. 2016. Grammatical error correction using neural machine translation. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 380–386, San Diego, California. Association for Co… find

↑ back to top