Elamur Research⌕
☰
home / papers / 2024 / Arabic Diacritics in the Wild: Exploiting Opportunities for Improved Diacritization

Arabic Diacritics in the Wild: Exploiting Opportunities for Improved Diacritization

Paper · 2024 · Salman Elgamal, Ossama Obeid, Mhd Kabbani et al.

Summary

Arabic is usually written without most of its optional diacritical marks, which makes the script highly ambiguous for NLP. Even so, writers sometimes add a few marks, roughly 1–2% of words in ordinary text, to steer readers toward the intended reading. The authors call these "Diacritics in the Wild" (WildDiacs) and ask whether such sparse hints can improve automatic diacritization.

They compare WildDiac usage across six genres: news, novels, children's books, poetry, UN political/legal documents and ChatGPT output. They also examine the diacritization conventions of the Penn Arabic Treebank and WikiNews. They define a stricter "maximal diacritization" target that includes word-initial and cross-word phonological effects. Two resources follow:

  • Wild2MaxDiacs: partially diacritized real-world words paired with their maximal diacritization in context.
  • WikiNewsMaxDiacs: a multi-reference revision of WikiNews.

On the modelling side, they extend the analyze-and-disambiguate pipeline in CAMeL Tools with a larger morphological database and contextual post-editing, so the system respects the hints. The result is large accuracy gains on both datasets. This shows that rare, author-supplied signals are worth exploiting rather than discarding.

Key points

  • The genre study spans news, novels, children's books, poetry, UN documents and ChatGPT output, which sometimes adds diacritics without being asked. In general news text, only about 1–2% of words carry any diacritic.
  • On Wild2MaxDiacs (All Dev), gains over the CAMeL Tools baseline break down as follows. No Wild2MaxDiacs data was used for training.
    • The extended morphological database adds 21.1%.
    • Contextual post-editing adds a further 10.2%.
    • All extensions together add 39.7%, reaching about 95.1% of Oracle accuracy.
  • On WikiNews, strict word-level accuracy from dediacritized input rose as follows:
    • Original WikiNews: from 47.1% to 84.5%.
    • Multi-reference WikiNewsMaxDiacs: from 47.5% to 89.7%.
    • About 64% of the latter gain comes from corrected base references and the rest from the added alternative references.
  • Analysis of all 354 Dev errors gives this breakdown:
    • 55% missing morphological analyses
    • 22% tagging failures
    • 11% wrong WildDiac input
    • 11% gold reference errors
    • 1% genuinely valid alternatives
  • Code and datasets are released open source (CAMeL-Lab/wild_diacritics).

Interesting ideas

  • I1. Use sparse, writer-provided hints as filters on the candidate analyses produced by a symbolic analyzer, before the statistical disambiguator picks one; a weak signal then becomes a hard constraint instead of noise.
  • I2. Separate the target form from what annotators habitually write: the paper defines a "maximal" standard and finds that existing gold sets miss phenomena like cross-word sound changes.
  • I3. Build multi-reference gold and report how much of the measured gain comes from fixing references versus accepting alternatives; this keeps "improvement" from being an artefact of a single narrow answer.
  • I4. Classify every error by severity, including explicit classes for bad input, gold mistakes and valid alternatives, and express scores relative to an oracle ceiling.
  • I5. Apply a context-sensitive post-editing pass across word boundaries, since a word's correct surface form depends on its neighbour (e.g., an epenthetic vowel before a definite article).

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (tokens): Signals already in the input, such as capitalization mid-sentence, hyphenation or explicit "a/an", could prune morphological and POS candidates before the tagger decides, keeping each choice traceable to the cue.
  • I2 (learn): Before trusting a treebank or silver set as reference, write down the exact target convention and audit the data against it, flagging gaps rather than training on them.
  • I3 (learn): Evaluation of parses or frames could store several acceptable gold variants and report fixed-reference and alternative-reference gains separately, as a guard against fooling ourselves.
  • I4 (learn): Error reports for the parser and readers could use such a taxonomy plus an oracle bound, showing whether the next fix belongs in the lexicon, the model, or the gold data.
  • I5 (back): Generation from the meaning graph could end with an explainable boundary-rule pass for neighbour-dependent forms like "a/an", contractions and elision.

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

🎓Mohamed bin Zayed University of Artificial IntelligenceAbu Dhabi · United Arab Emirates138 works · 6,590 citations🎓American University of SharjahSharjah · United Arab Emirates4 works · 81 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 works

🧭 Topics 1Research subjects and languages the work is about.Semitic languages · 3,066 works

Cite

Cite as Salman Elgamal, Ossama Obeid, Mhd Kabbani, Go Inoue and Nizar Habash. 2024. Arabic Diacritics in the Wild: Exploiting Opportunities for Improved Diacritization. Proceedings of the 62nd 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. DOI 10.18653/v1/2024.acl-long.792

@inproceedings{elgamal2024arabic,
  title = {{Arabic Diacritics in the Wild: Exploiting Opportunities for Improved Diacritization}},
  author = {Salman Elgamal, Ossama Obeid, Mhd Kabbani, Go Inoue and Nizar Habash},
  booktitle = {Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2024},
  doi = {10.18653/v1/2024.acl-long.792},
  url = {https://doi.org/10.18653/v1/2024.acl-long.792}
}

References

59 references, 40 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Gheith A. Abandah, Alex Graves, Balkees Al-Shagoor, Alaa Arabiyat, Fuad Jamour, and Majid Al-Taee. 2015. Automatic diacritization of Arabic text using recurrent neural networks. International Journal on Document Analysis and Recognition (IJDAR), 18(2):183–197. ↗
  2. Ahmed Abdelali, Mohammed Attia, Younes Samih, Kareem Darwish, and Hamdy Mubarak. 2019. Diacritization of Maghrebi Arabic sub-dialects. ↗
  3. Ahmed Abdelali, Nadir Durrani, Cenk Demiroglu, Fahim Dalvi, Hamdy Mubarak, and Kareem Darwish. 2022. NatiQ: An end-to-end text-to-speech system for Arabic. In Proceedings of the The Seventh Arabic Natural Language Processing Workshop (WANLP), pages 394–398, Abu Dhabi, United Arab Emirates (Hybrid).
  4. Muhamed Al Khalil, Hind Saddiki, Nizar Habash, and Latifa Alfalasi. 2018. A Leveled Reading Corpus of Modern Standard Arabic. In Proceedings of the Language Resources and Evaluation Conference (LREC), Miyazaki, Japan.
  5. Nouf Alabbasi, Mohamed Al-Badrashiny, Maryam Aldahmani, Ahmed AlDhanhani, Abdullah Saleh Alhashmi, Fawaghy Ahmed Alhashmi, Khalid Al Hashemi, Rama Emad Alkhobbi, Shamma T Al Maazmi, Mohammed Ali Alyafeai, Mariam M Alzaabi, Mohamed Saqer Alzaabi, Fatma Khalid Badri, Kareem Darwish, Ehab Mansour Diab, Muhammad Morsy Elmallah, Amira Ayman Elnashar, Ashraf Hatim Elneima, MHD Tameem Kabbani, Nour Rabih, Ahmad Saad, and Ammar Mamoun Sousou. 2022. Gulf Arabic diacritization: Guidelines, initial dataset, and results. In Proceedings of the The Seventh Arabic Natural Language Processing Workshop (WANLP), pages 356– 360, Abu Dhabi, United Arab Emirates (Hybrid). Association for Computational Linguistics.
  6. Hanan Aldarmaki and Ahmad Ghannam. 2023. Diacritic Recognition Performance in Arabic ASR. In Proc. INTERSPEECH 2023, pages 361–365.
  7. Bashar Alhafni, Go Inoue, Christian Khairallah, and Nizar Habash. 2023. Advancements in Arabic grammatical error detection and correction: An empirical investigation. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 6430–6448, Singapore.
  8. Badr AlKhamissi, Muhammad ElNokrashy, and Mohamed Gabr. 2020. Deep diacritization: Efficient hierarchical recurrence for improved Arabic diacritization. In Proceedings of the Fifth Arabic Natural Language Processing Workshop, pages 38–48, Barcelona, Spain (Online). ↗
  9. Sawsan Alqahtani, Hanan Aldarmaki, and Mona Diab. 2019. Homograph disambiguation through selective diacritic restoration. In Proceedings of the Fourth Arabic Natural Language Processing Workshop, pages 49–59, Florence, Italy. find
  10. Sawsan Alqahtani, Mahmoud Ghoneim, and Mona Diab. 2016. Investigating the impact of various partial diacritization schemes on Arabic-English statistical machine translation. In Conferences of the Association for Machine Translation in the Americas: MT Researchers’ Track, pages 191–204, Austin, TX, USA. The Association for Machine Translation in the Americas. find
  11. Sawsan Alqahtani, Ajay Mishra, and Mona Diab. 2020. A multitask learning approach for diacritic restoration. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 8238–8247, Online.
  12. Parnia Bahar, Mattia Di Gangi, Nick Rossenbach, and Mohammad Zeineldeen. 2023. Take the Hint: Improving Arabic Diacritization with Partially-Diacritized Text. In Proc. INTERSPEECH 2023, pages 3949–3953.
  13. Yonatan Belinkov and James Glass. 2015. Arabic diacritization with recurrent neural networks. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 2281– 2285, Lisbon, Portugal.
  14. Kareem Darwish, Ahmed Abdelali, Hamdy Mubarak, and Mohamed Eldesouki. 2021. Arabic diacritic recovery using a feature-rich bilstm model. ACM Trans. Asian Low-Resour. Lang. Inf. Process., 20(2). ↗
  15. Kareem Darwish, Hamdy Mubarak, and Ahmed Abdelali. 2017. Arabic diacritization: Stats, rules, and hacks. In Proceedings of the Third Arabic Natural Language Processing Workshop, pages 9–17.
  16. Mona Diab, Mahmoud Ghoneim, and Nizar Habash. 2007. Arabic diacritization in the context of statistical machine translation. In Proceedings of Machine Translation Summit XI: Papers, Copenhagen, Denmark.
  17. Mona Diab, Nizar Habash, Reem Faraj, and May Ah-mar. 2010. Guidelines for the creation of resources for Arabic dialects. In Proceedings of the Workshop on Language Resources and Human Language Technology for Semitic Languages. find
  18. Mahmoud El-Haj. 2020. Habibi - a multi dialect multi national Arabic song lyrics corpus. In Proceedings of the Twelfth Language Resources and Evaluation Conference, pages 1318–1326, Marseille, France. European Language Resources Association.
  19. AbdelRahim Elmadany, El Moatez Billah Nagoudi, and Muhammad Abdul-Mageed. 2023. Octopus: A multitask model and toolkit for Arabic natural language generation. In Proceedings of ArabicNLP 2023, pages 232–243, Singapore (Hybrid).
  20. Muhammad Morsy Elmallah, Mahmoud Reda, Kareem Darwish, Abdelrahman El-Sheikh, Ashraf Hatim El-neima, Murtadha Aljubran, Nouf Alsaeed, Reem Mohammed, and Mohamed Al-Badrashiny. 2024. Arabic diacritization using morphologically informed character-level model. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pages 1446–1454, Torino, Italia. ELRA and ICCL.
  21. Muhammad ElNokrashy and Badr AlKhamissi. 2024. Partial diacritization: A context-contrastive inference approach. find
  22. Fadhl Eryani and Nizar Habash. 2021. Automatic Romanization of Arabic bibliographic records. In Proceedings of the Sixth Arabic Natural Language Processing Workshop, pages 213–218, Kyiv, Ukraine (Virtual). ↗
  23. Saeed Esmail, Kfir Bar, and Nachum Dershowitz. 2022. How much does lookahead matter for disambiguation? partial Arabic diacritization case study. Computational Linguistics, 48(4):1103–1123. ↗
  24. Ali Fadel, Ibraheem Tuffaha, Bara’ Al-Jawarneh, and Mahmoud Al-Ayyoub. 2019a. Arabic text diacritization using deep neural networks.
  25. Ali Fadel, Ibraheem Tuffaha, Bara’ Al-Jawarneh, and Mahmoud Al-Ayyoub. 2019b. Neural Arabic text diacritization: State of the art results and a novel approach for machine translation. In Proceedings of the 6th Workshop on Asian Translation, pages 215– 225, Hong Kong, China. ↗
  26. David Graff, Mohamed Maamouri, Basma Bouziri, Sondos Krouna, Seth Kulick, and Tim Buckwalter. 2009. Standard Arabic Morphological Analyzer (SAMA) Version 3.1. Linguistic Data Consortium LDC2009E73. find
  27. Nizar Habash and Owen Rambow. 2005. Arabic tokenization, part-of-speech tagging and morphological disambiguation in one fell swoop. In Proceedings of the Conference of the Association for Computational Linguistics (ACL), pages 573–580, Ann Arbor, Michigan.
  28. Nizar Habash and Owen Rambow. 2007. Arabic diacritization through full morphological tagging. In Human Language Technologies 2007: The Conference of the North American Chapter of the Association for Computational Linguistics; Companion Volume, Short Papers, pages 53–56, Rochester, New York.
  29. Nizar Habash, Anas Shahrour, and Muhamed Al-Khalil. 2016. Exploiting Arabic diacritization for high quality automatic annotation. In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC’16), pages 4298–4304, Portorož, Slovenia. European Language Resources Association (ELRA).
  30. Nizar Habash, Abdelhadi Soudi, and Tim Buckwalter. 2007. On Arabic Transliteration. In A. van den Bosch and A. Soudi, editors, Arabic Computational Morphology: Knowledge-based and Empirical Methods, pages 15–22. Springer, Netherlands. ↗
  31. Nizar Y Habash. 2010. Introduction to Arabic natural language processing, volume 3. Morgan & Claypool Publishers. find
  32. Nawar Halabi. 2016. Modern standard Arabic phonetics for speech synthesis. Ph.D. thesis, University of Southampton. find
  33. Osama Hamed and Torsten Zesch. 2018. The role of diacritics in increasing the difficulty of Arabic lexical recognition tests. In Proceedings of the 7th workshop on NLP for Computer Assisted Language Learning, pages 23–31, Stockholm, Sweden. LiU Electronic Press.
  34. Go Inoue, Salam Khalifa, and Nizar Habash. 2022. Morphosyntactic tagging with pre-trained language models for Arabic and its dialects. In Findings of the Association for Computational Linguistics: ACL 2022, pages 1708–1719, Dublin, Ireland.
  35. Mustafa Jarrar, Nizar Habash, Faeq Alrimawi, Diyam Akra, and Nasser Zalmout. 2016. Curras: an annotated corpus for the Palestinian Arabic dialect. Language Resources and Evaluation, pages 1–31. ↗
  36. Mustafa Jarrar, Fadi Zaraket, Rami Asia, and Hamzeh Amayreh. 2018. Diacritic-based matching of Arabic words. ACM Trans. Asian Low-Resour. Lang. Inf. Process., 18(2). ↗
  37. Mohamed Maamouri, Ann Bies, Tim Buckwalter, and Wigdan Mekki. 2004. The Penn Arabic Treebank: Building a Large-Scale Annotated Arabic Corpus. In Proceedings of the International Conference on Arabic Language Resources and Tools, pages 102– 109, Cairo, Egypt. find
  38. Mokthar Ali Hasan Madhfar and Ali Mustafa Qamar. 2021. Effective deep learning models for automatic diacritization of Arabic text. IEEE Access, 9:273– 288. ↗
  39. Hamdy Mubarak, Ahmed Abdelali, Hassan Sajjad, Younes Samih, and Kareem Darwish. 2019. Highly effective Arabic diacritization using sequence to sequence modeling. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 2390–2395, Minneapolis, Minnesota.
  40. Ossama Obeid, Go Inoue, and Nizar Habash. 2022. Camelira: An Arabic multi-dialect morphological disambiguator. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 319–326, Abu Dhabi, UAE.
  41. Ossama Obeid, Nasser Zalmout, Salam Khalifa, Dima Taji, Mai Oudah, Bashar Alhafni, Go Inoue, Fadhl Eryani, Alexander Erdmann, and Nizar Habash. 2020. CAMeL tools: An open source python toolkit for Arabic natural language processing. In Proceedings of the Twelfth Language Resources and Evaluation Conference, pages 7022–7032, Marseille, France.
  42. Robert Parker, David Graff, Ke Chen, Junbo Kong, and Kazuaki Maeda. 2011. Arabic Gigaword Fifth Edition. LDC catalog number No. LDC2011T11, ISBN 1-58563-595-2. ↗
  43. Arfath Pasha, Mohamed Al-Badrashiny, Mona Diab, Ahmed El Kholy, Ramy Eskander, Nizar Habash, Manoj Pooleery, Owen Rambow, and Ryan Roth. 2014. Madamira: A fast, comprehensive tool for morphological analysis and disambiguation of Arabic. In Proceedings of the Language Resources and Evaluation Conference (LREC), pages 1094–1101, Reykjavik, Iceland.
  44. Roozbeh Pournader, Bob Hallissy, and Lorna Evans. 2021. Unicode Arabic Mark Rendering. find
  45. Han Qin, Guimin Chen, Yuanhe Tian, and Yan Song. 2021. Improving Arabic diacritization with regularized decoding and adversarial training. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 2: Short Papers), pages 534–542, Online.
  46. Ryan Roth, Owen Rambow, Nizar Habash, Mona Diab, and Cynthia Rudin. 2008. Arabic morphological tagging, diacritization, and lemmatization using lexeme models and feature ranking. In Proceedings of ACL-08: HLT, Short Papers, pages 117–120, Columbus, Ohio.
  47. Dima Taji, Salam Khalifa, Ossama Obeid, Fadhl Eryani, and Nizar Habash. 2018. An Arabic morphological analyzer and generator with copious features. In Proceedings of the Fifteenth Workshop on Computational Research in Phonetics, Phonology, and Morphology, pages 140–150, Brussels, Belgium.
  48. Brian Thompson and Ali Alshehri. 2022. Improving Arabic diacritization by learning to diacritize and translate. In Proceedings of the 19th International Conference on Spoken Language Translation (IWSLT 2022), pages 11–21, Dublin, Ireland (in-person and online). Association for Computational Linguistics.
  49. Catalin Ungurean, Dragos Burileanu, Vladimir Popescu, Cristian Negrescu, and Aurelian Dervis. 2008. Automatic diacritic restoration for a TTS-based e-mail reader application. UPB Scientific Bulletin, Series C, 70(4):3–12. find
  50. Ken Whistler. 2023a. Unicode Character Database. find
  51. Ken Whistler. 2023b. Unicode Normalization Forms. find
  52. Waleed A. Yousef, Omar M. Ibrahime, Taha M. Mad-bouly, and Moustafa A. Mahmoud. 2019. Learning meters of Arabic and English poems with recurrent neural networks: a step forward for language understanding and synthesis. ↗
  53. Nasser Zalmout and Nizar Habash. 2017. Don’t throw those morphological analyzers away just yet: Neural morphological disambiguation for Arabic. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 704–713, Copenhagen, Denmark.
  54. Nasser Zalmout and Nizar Habash. 2019. Adversarial multitask learning for joint multi-feature and multi-dialect morphological modeling. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 1775–1786, Florence, Italy.
  55. Nasser Zalmout and Nizar Habash. 2020. Joint diacritization, lemmatization, normalization, and fine-grained morphological tagging. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 8297–8307, Online.
  56. Taha Zerrouki and Amar Balla. 2017. Tashkeela: Novel corpus of Arabic vocalized texts, data for auto-diacritization systems. Data in brief, 11:147–151. find
  57. Michal Ziemski, Marcin Junczys-Dowmunt, and Bruno Pouliquen. 2016. The united nations parallel corpus v1.0. In Proceedings of the Language Resources and Evaluation Conference (LREC), Portorož, Slovenia.
  58. Imed Zitouni, Jeffrey S. Sorensen, and Ruhi Sarikaya. 2006. Maximum entropy based restoration of Arabic diacritics. In Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, pages 577–584, Sydney, Australia.
  59. Anas Shahrour, Salam Khalifa, and Nizar Habash. 2015. Improving Arabic diacritization through syntactic analysis. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 1309–1315, Lisbon, Portugal.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/elgamal-2024-arabic-diacritics-wild-exploiting-opportunities/