---
title: "TURJUMAN — paper by Nagoudi et al. (2022)"
description: "The toolkit offers four decoding strategies: greedy, beam, top-k, and nucleus (top-p) sampling. 48 references, cited by 14 papers here."
authors: ["El Moatez Billah Nagoudi", "AbdelRahim Elmadany", "Muhammad Abdul-Mageed"]
year: 2022
venue: "Proceedinsg of the 5th Workshop on Open-Source Arabic Corpora and Processing Tools with Shared Tasks on Qur'an QA and Fine-Grained Hate Speech Detection"
source: "https://aclanthology.org/2022.osact-1.1/"
licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)"
licence_url: "https://creativecommons.org/licenses/by-nc/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/nagoudi-2022-turjuman-public-toolkit-neural-arabic/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/nagoudi-2022-turjuman-public-toolkit-neural-arabic/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2022](https://md.research.elamur.ai/free/en/papers/2022/page/6/) / TURJUMAN: A Public Toolkit for Neural Arabic Machine Translation

# TURJUMAN: A Public Toolkit for Neural Arabic Machine Translation

**Authors:** [El Moatez Billah Nagoudi](https://md.research.elamur.ai/free/en/authors/el-moatez-billah-nagoudi/), [AbdelRahim Elmadany](https://md.research.elamur.ai/free/en/authors/abdelrahim-elmadany/) and [Muhammad Abdul-Mageed](https://md.research.elamur.ai/free/en/authors/muhammad-abdul-mageed/) · **Year:** 2022 · **Venue:** [Proceedinsg of the 5th Workshop on Open-Source Arabic Corpora and Processing Tools with Shared Tasks on Qur'an QA and Fine-Grained Hate Speech Detection](https://md.research.elamur.ai/free/en/venues/osact/) ([Workshop on Open-Source Arabic Corpora and Processing Tools with Shared Tasks 2022](https://md.research.elamur.ai/free/en/venues/osact/2022/)) · **Publisher:** [European Language Resources Association (ELRA)](https://md.research.elamur.ai/free/en/publishers/european-language-resources-association-elra/) · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2022.osact-1.1/) · [PDF](https://aclanthology.org/2022.osact-1.1.pdf) · [OpenAlex](https://openalex.org/W4281717508)

**Citations:** 7 citations (OpenAlex, 2026-10-14) · [14](https://md.research.elamur.ai/free/en/papers/nagoudi-2022-turjuman-public-toolkit-neural-arabic/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
Arabic NLP already had public tools for morphology and social-meaning detection, but there were few openly available machine translation tools. The authors address this with TURJUMAN, a Python toolkit that translates from 20 source languages into Modern Standard Arabic (MSA). It is built by fine-tuning AraT5, an Arabic text-to-text Transformer. For training and evaluation they assemble AraOPUS-20, a parallel benchmark drawn from OPUS. Because OPUS contains many misaligned or noisy pairs, they score each pair with a multilingual SBERT sentence encoder and keep only pairs that are semantically close. The toolkit offers four decoding strategies: greedy, beam, top-k, and nucleus (top-p) sampling. Sampling several outputs for one source sentence also makes it a source of Arabic paraphrases. AraT5 is compared with baselines on the new benchmark, but the score tables are not in the excerpt we received. The authors present TURJUMAN as an extensible baseline rather than a state-of-the-art system. Its value is that the tool, the cleaned benchmark and its standard splits are all public.

## Key points
- TURJUMAN translates from 20 languages into MSA by fine-tuning AraT5. It is released at github.com/UBC-NLP/turjuman together with the AraOPUS-20 benchmark.
- AraOPUS-20 is extracted from OPUS, which has over 2.7 billion parallel sentences in 90 languages. It is released in standard splits.
- The 20 languages range from high-resource ones (English, French, Spanish) to low-resource ones (Cebuano, Tamashek, Yoruba).
- Noise filtering uses multilingual SBERT similarity between source and Arabic sentences. In the paper's examples, low-similarity pairs (some scoring below zero) are clearly wrong translations.
- The authors deliberately did not train on all the parallel data available to them. They list resources for future extension, including the UN corpus, IWSLT, Arab-Acquis, MADAR, APT, PADIC and Dial2MSA.

## Interesting ideas
- Filtering a scraped parallel corpus by cross-lingual embedding similarity is a cheap, language-agnostic way to drop misaligned pairs before training.
- Showing a table of sample pairs with their scores and keep/drop decisions makes the cleaning threshold easy to audit by eye.
- Running several decoding strategies on one input turns a translator into a paraphrase generator at no extra training cost.
- Releasing a deliberately modest, extensible baseline with fixed splits, rather than chasing top scores, makes later comparisons fair and reproducible.

## For Mova

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

- **I1** (learn): Silver data for translation or cross-lingual frames could be screened the same way, with the similarity score kept as an explainable reason for each rejection.
- **I2** (learn): Each silver-data filter could publish a small scored sample of accepted and rejected items, acting as a human-readable check on the filter.
- **I3** (back): The generator from the meaning graph could produce several surface variants per graph; checking that all of them map back to the same graph would test the compressor.
- **I4** (answer): School-test and fable-QA evaluation sets could be frozen into standard splits, so future reader versions are measured against the same fixed baseline.

*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

- [University of British Columbia](https://md.research.elamur.ai/free/en/universities/university-of-british-columbia/) — University, Vancouver, Canada · 104 works · 37,403 citations

📚 Shelves (4): [🖥️ 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

🧭 Topics (2): [🔁 Machine translation](https://md.research.elamur.ai/free/en/topics/machine-translation/) · 1,105 works · [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 3,066 works

## Cite

Cite as: El Moatez Billah Nagoudi, AbdelRahim Elmadany and Muhammad Abdul-Mageed. 2022. TURJUMAN: A Public Toolkit for Neural Arabic Machine Translation. Proceedinsg of the 5th Workshop on Open-Source Arabic Corpora and Processing Tools with Shared Tasks on Qur'an QA and Fine-Grained Hate Speech Detection.

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

```bibtex
@inproceedings{nagoudi2022turjuman,
  title = {{TURJUMAN: A Public Toolkit for Neural Arabic Machine Translation}},
  author = {El Moatez Billah Nagoudi, AbdelRahim Elmadany and Muhammad Abdul-Mageed},
  booktitle = {Proceedinsg of the 5th Workshop on Open-Source Arabic Corpora and Processing Tools with Shared Tasks on Qur'an QA and Fine-Grained Hate Speech Detection},
  year = {2022},
  url = {https://aclanthology.org/2022.osact-1.1/}
}
```

- [← previous The Arabic Parallel Gender Corpus 2.0: Extensions and Analyses](https://md.research.elamur.ai/free/en/papers/alhafni-2022-arabic-parallel-gender-corpus-2/)
- [next → A Double-Graph Based Framework for Frame Semantic Parsing](https://md.research.elamur.ai/free/en/papers/zheng-2022-double-graph-based-framework-frame/)
- [↑ back to papers of 2022](https://md.research.elamur.ai/free/en/papers/2022/page/6/)

## References

48 references, 22 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Abdul-Mageed, M., Zhang, C., Hashemi, A., and Nagoudi, E. M. B. (2019). [AraNet: A Deep Learning Toolkit for Arabic Social Media](https://md.research.elamur.ai/free/en/papers/mageed-2020-aranet-deep-learning-toolkit-arabic/). arXiv preprint arXiv:1912.13072. [↗](https://arxiv.org/abs/1912.13072)
2. Akeel, M. and Mishra, R. (2014). Ann and rule based method for english to arabic machine translation. Int. Arab J. Inf. Technol., 11(4):396–405. [find](https://scholar.google.com/scholar?q=Akeel%2C+M%2E+and+Mishra%2C+R%2E+%282014%29%2E+Ann+and+rule+based+method+for+english+to+arabic+machine+translation%2E+Int%2E+Arab+J%2E+Inf%2E+Technol%2E%2C+11%284%29%3A396%E2%80%93405%2E)
3. Almahairi, A., Cho, K., Habash, N., and Courville, A. (2016). [First Result on Arabic Neural Machine Translation](https://md.research.elamur.ai/free/en/papers/almahairi-2016-first-result-arabic-neural-machine/). arXiv preprint arXiv:1606.02680. [↗](https://arxiv.org/abs/1606.02680)
4. Alrajeh, A. (2018). A Recipe for Arabic-English Neural Machine Translation. arXiv preprint arXiv:1808.06116. [↗](https://arxiv.org/abs/1808.06116)
5. Aqlan, F., Fan, X., Alqwbani, A., and Al-Mansoub, A. (2019). Arabic–chinese neural machine translation: Romanized arabic as subword unit for arabic-sourced translation. IEEE Access, 7:133122–133135. [↗](https://doi.org/10.1109/access.2019.2941161)
6. Bahdanau, D., Cho, K., and Bengio, Y. (2014). [Neural machine translation by jointly learning to align and translate](https://md.research.elamur.ai/free/en/papers/bahdanau-2014-neural-machine-translation-jointly-learning/). arXiv preprint arXiv:1409.0473. [↗](https://arxiv.org/abs/1409.0473)
7. Bakr, H. A., Shaalan, K., and Ziedan, I. (2008). A hybrid approach for converting written egyptian colloquial dialect into diacritized arabic. In The 6th international conference on informatics and systems, infos2008. cairo university. [find](https://scholar.google.com/scholar?q=Bakr%2C+H%2E+A%2E%2C+Shaalan%2C+K%2E%2C+and+Ziedan%2C+I%2E+%282008%29%2E+A+hybrid+approach+for+converting+written+egyptian+colloquial+dialect+into+diacritized+arabic%2E+In+The+6th+international+conference+on+informatics+and+sy)
8. Baniata, L. H., Park, S., and Park, S.-B. (2018). A neural machine translation model for arabic dialects that utilizes multitask learning (mtl). Computational intelligence and neuroscience, 2018. [↗](https://doi.org/10.1155/2018/7534712)
9. Bouamor, H., Habash, N., and Oflazer, K. (2014). [A multidialectal parallel corpus of arabic](https://md.research.elamur.ai/free/en/papers/bouamor-2014-multidialectal-parallel-corpus-arabic/). In LREC, pages 1240–1245.
10. Bouamor, H., Habash, N., Salameh, M., Zaghouani, W., Rambow, O., Abdulrahim, D., Obeid, O., Khalifa, S., Eryani, F., Erdmann, A., et al. (2018). The madar arabic dialect corpus and lexicon. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC-2018). [find](https://scholar.google.com/scholar?q=Bouamor%2C+H%2E%2C+Habash%2C+N%2E%2C+Salameh%2C+M%2E%2C+Zaghouani%2C+W%2E%2C+Rambow%2C+O%2E%2C+Abdulrahim%2C+D%2E%2C+Obeid%2C+O%2E%2C+Khalifa%2C+S%2E%2C+Eryani%2C+F%2E%2C+Erdmann%2C+A%2E%2C+et+al%2E+%282018%29%2E+The+madar+arabic+dialect+corpus+and+lexicon%2E+In+Proceed)
11. Cettolo, M., Niehues, J., Stüker, S., Bentivogli, L., and Federico, M. (2013). Report on the 10th iwslt evaluation campaign. In Proceedings of the International Workshop on Spoken Language Translation, Heidelberg, Germany. https://www.computecanada.ca https://arc.ubc.ca/ubc-arc-sockeye [↗](https://doi.org/10.5445/ir/1000145021)
12. Cettolo, M., Jan, N., Sebastian, S., Bentivogli, L., Cattoni, R., and Federico, M. (2016). The iwslt 2016 evaluation campaign. In International Workshop on Spoken Language Translation. [find](https://scholar.google.com/scholar?q=Cettolo%2C+M%2E%2C+Jan%2C+N%2E%2C+Sebastian%2C+S%2E%2C+Bentivogli%2C+L%2E%2C+Cattoni%2C+R%2E%2C+and+Federico%2C+M%2E+%282016%29%2E+The+iwslt+2016+evaluation+campaign%2E+In+International+Workshop+on+Spoken+Language+Translation%2E)
13. Cettolo, M., Federico, M., Bentivogli, L., Jan, N., Sebastian, S., Katsuitho, S., Koichiro, Y., and Christian, F. (2017). [Overview of the iwslt 2017 evaluation campaign](https://md.research.elamur.ai/free/en/papers/cettolo-2017-overview-iwslt-2017-evaluation-campaign/). In International Workshop on Spoken Language Translation, pages 2–14.
14. Cormen, T. H., Leiserson, C. E., Rivest, R. L., and Stein, C. (2009). Introduction to algorithms. [find](https://scholar.google.com/scholar?q=Cormen%2C+T%2E+H%2E%2C+Leiserson%2C+C%2E+E%2E%2C+Rivest%2C+R%2E+L%2E%2C+and+Stein%2C+C%2E+%282009%29%2E+Introduction+to+algorithms%2E)
15. Darwish, K. and Mubarak, H. (2016). [Farasa: A new fast and accurate Arabic word segmenter](https://md.research.elamur.ai/free/en/papers/darwish-2016-farasa-new-fast-accurate-arabic/). In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC’16), pages 1070–1074, Portorož, Slovenia, May. European Language Resources Association (ELRA).
16. Durrani, N., Dalvi, F., Sajjad, H., and Vogel, S. (2017). Qcri machine translation systems for iwslt 16. arXiv preprint arXiv:1701.03924. [↗](https://arxiv.org/abs/1701.03924)
17. El-Kahlout, I. D., Bektaş, E., Erdem, N. Ş., and Kaya, H. (2019). Translating between morphologically rich languages: An arabic-to-turkish machine translation system. In Proceedings of the Fourth Arabic Natural Language Processing Workshop, pages 158–166. [find](https://scholar.google.com/scholar?q=El%2DKahlout%2C+I%2E+D%2E%2C+Bektas%CC%A7%2C+E%2E%2C+Erdem%2C+N%2E+S%CC%A7%2E%2C+and+Kaya%2C+H%2E+%282019%29%2E+Translating+between+morphologically+rich+languages%3A+An+arabic%2Dto%2Dturkish+machine+translation+system%2E+In+Proceedings+of+the+Fourth+Ar)
18. Elmahdy, M., Hasegawa-Johnson, M., and Mustafawi, E. (2014). [Development of a tv broadcasts speech recognition system for qatari arabic](https://md.research.elamur.ai/free/en/papers/elmahdy-2014-development-tv-broadcasts-speech-recognition/). In LREC, pages 3057– 3061.
19. Fadaee, M., Bisazza, A., and Monz, C. (2017). [Data augmentation for low-resource neural machine translation](https://md.research.elamur.ai/free/en/papers/fadaee-2017-data-augmentation-low-resource-neural/). arXiv preprint arXiv:1705.00440. [↗](https://arxiv.org/abs/1705.00440)
20. Fan, A., Lewis, M., and Dauphin, Y. (2018). [Hierarchical neural story generation](https://md.research.elamur.ai/free/en/papers/fan-2018-hierarchical-neural-story-generation/). arXiv preprint arXiv:1805.04833. [↗](https://arxiv.org/abs/1805.04833)
21. Farha, I. A. and Magdy, W. (2019). [Mazajak: An online arabic sentiment analyser](https://md.research.elamur.ai/free/en/papers/farha-2019-mazajak-online-arabic-sentiment-analyser/). In Proceedings of the Fourth Arabic Natural Language Processing Workshop, pages 192–198.
22. Farhan, W., Talafha, B., Abuammar, A., Jaikat, R., Al-Ayyoub, M., Tarakji, A. B., and Toma, A. (2020). Unsupervised dialectal neural machine translation. Information Processing & Management, 57(3):102181. [↗](https://doi.org/10.1016/j.ipm.2019.102181)
23. Federico, M., Cettolo, M., Bentivogli, L., Michael, P., and Sebastian, S. (2012). Overview of the iwslt 2012 evaluation campaign. In IWSLT-International Workshop on Spoken Language Translation, pages 12–33. [find](https://scholar.google.com/scholar?q=Federico%2C+M%2E%2C+Cettolo%2C+M%2E%2C+Bentivogli%2C+L%2E%2C+Michael%2C+P%2E%2C+and+Sebastian%2C+S%2E+%282012%29%2E+Overview+of+the+iwslt+2012+evaluation+campaign%2E+In+IWSLT%2DInternational+Workshop+on+Spoken+Language+Translation%2C+pages+)
24. Ghoneim, M. and Diab, M. (2013). Multiword expressions in the context of statistical machine translation. In Proceedings of the Sixth International Joint Conference on Natural Language Processing, pages 1181–1187. [↗](https://doi.org/10.13140/2.1.4636.8965)
25. Guellil, I., Azouaou, F., and Abbas, M. (2017). Neural vs statistical translation of algerian arabic dialect written with arabizi and arabic letter. In The 31st Pacific Asia Conference on Language, Information and Computation PACLIC, volume 31, page 2017. [find](https://scholar.google.com/scholar?q=Guellil%2C+I%2E%2C+Azouaou%2C+F%2E%2C+and+Abbas%2C+M%2E+%282017%29%2E+Neural+vs+statistical+translation+of+algerian+arabic+dialect+written+with+arabizi+and+arabic+letter%2E+In+The+31st+Pacific+Asia+Conference+on+Language%2C+In)
26. Habash, N. and Hu, J. (2009). Improving arabic-chinese 9 statistical machine translation using english as pivot language. In Proceedings of the Fourth Workshop on Statis-Online, June. Association for Computational Linguistics. tical Machine Translation, pages 173–181. Nagoudi, E. M. B., Elmadany, A., and Abdul-Mageed, M. [find](https://scholar.google.com/scholar?q=Habash%2C+N%2E+and+Hu%2C+J%2E+%282009%29%2E+Improving+arabic%2Dchinese+9+statistical+machine+translation+using+english+as+pivot+language%2E+In+Proceedings+of+the+Fourth+Workshop+on+Statis%2DOnline%2C+June%2E+Association+for+)
27. Habash, N., Zalmout, N., Taji, D., Hoang, H., and Alzate, (2022). Arat5: Text-to-text transformers for arabic lan-M. (2017). A parallel corpus for evaluating machine guage generation. Online, May. Association for Compu-translation between Arabic and European languages. In tational Linguistics. Proceedings of the 15th Conference of the European Noll, E., Oudah, M., and Habash, N. (2019). Simple auto-Chapter of the Association for Computational Linguis-matic post-editing for arabic-japanese machine transla-tics: Volume 2, Short Papers, pages 235–241, Valencia, tion. arXiv preprint arXiv:1907.06210. Spain, April. Association for Computational Linguistics. [↗](https://arxiv.org/abs/1907.06210)
28. Obeid, O., Zalmout, N., Khalifa, S., Taji, D., Oudah, M., Holtzman, A., Buys, J., Du, L., Forbes, M., and Choi, Y. Alhafni, B., Eryani, F., Erdmann, A., and Habash, N. (2019). [The curious case of neural text degeneration](https://md.research.elamur.ai/free/en/papers/holtzman-2019-curious-case-neural-text-degeneration/). (2020a). Camel tools: An open source python toolkit for arXiv preprint arXiv:1904.09751. arabic natural language processing. 05. [↗](https://arxiv.org/abs/1904.09751)
29. Junczys-Dowmunt, M., Dwojak, T., and Hoang, H. (2016). Obeid, O., Zalmout, N., Khalifa, S., Taji, D., Oudah, M., Is neural machine translation ready for deployment. A Alhafni, B., Inoue, G., Eryani, F., Erdmann, A., and case study on, 30. [find](https://scholar.google.com/scholar?q=Junczys%2DDowmunt%2C+M%2E%2C+Dwojak%2C+T%2E%2C+and+Hoang%2C+H%2E+%282016%29%2E+Obeid%2C+O%2E%2C+Zalmout%2C+N%2E%2C+Khalifa%2C+S%2E%2C+Taji%2C+D%2E%2C+Oudah%2C+M%2E%2C+Is+neural+machine+translation+ready+for+deployment%2E+A+Alhafni%2C+B%2E%2C+Inoue%2C+G%2E%2C+Eryani%2C+F)
30. Habash, N. (2020b). CAMeL tools: An open source Koehn, P. (2009). Statistical machine translation. Cam-python toolkit for Arabic natural language processing. In bridge University Press. Proceedings of the 12th Language Resources and Evaluation Conference, pages 7022–7032, Marseille, France, Lample, G., Ott, M., Conneau, A., Denoyer, L., and Ran-May. European Language Resources Association. zato, M. (2018). Phrase-based & neural unsupervised machine translation. arXiv preprint arXiv:1804.07755. [↗](https://arxiv.org/abs/1804.07755)
31. Ott, M., Edunov, S., Baevski, A., Fan, A., Gross, S., Ng, N., Grangier, D., and Auli, M. (2019). [fairseq: A fast, Lewis, M., Liu, Y., Goyal, N., Ghazvininejad, M., Mo-extensible toolkit for sequence modeling](https://md.research.elamur.ai/free/en/papers/ott-2019-fairseq-fast-extensible-toolkit-sequence/). arXiv preprint hamed, A., Levy, O., Stoyanov, V., and Zettlemoyer, arXiv:1904.01038. [↗](https://arxiv.org/abs/1904.01038)
32. L. (2020). Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and Pasha, A., Al-Badrashiny, M., Diab, M., El Kholy, A., comprehension. In Proceedings of the 58th Annual Eskander, R., Habash, N., Pooleery, M., Rambow, O., Meeting of the Association for Computational Linguis-and Roth, R. (2014). MADAMIRA: A fast, compre-tics, pages 7871–7880. hensive tool for morphological analysis and disambiguation of Arabic. In Proceedings of the Ninth Interna-Lhoest, Q., del Moral, A. V., Jernite, Y., Thakur, A., von tional Conference on Language Resources and Eval-Platen, P., Patil, S., Chaumond, J., Drame, M., Plu, J., uation (LREC’14), pages 1094–1101, Reykjavik, Ice-Tunstall, L., et al. (2021). Datasets: A community liland, May. European Language Resources Association brary for natural language processing. In Proceedings (ELRA). of the 2021 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., 175–184. [find](https://scholar.google.com/scholar?q=L%2E+%282020%29%2E+Bart%3A+Denoising+sequence%2Dto%2Dsequence+pre%2Dtraining+for+natural+language+generation%2C+translation%2C+and+Pasha%2C+A%2E%2C+Al%2DBadrashiny%2C+M%2E%2C+Diab%2C+M%2E%2C+El+Kholy%2C+A%2E%2C+comprehension%2E+In+Proceedings+of+th)
33. Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al. (2019). Pytorch: An imperative style, high-Liu, Y., Gu, J., Goyal, N., Li, X., Edunov, S., Ghazvinine-performance deep learning library. Advances in neural jad, M., Lewis, M., and Zettlemoyer, L. (2020). Multi-information processing systems, 32:8026–8037. lingual denoising pre-training for neural machine translation. Transactions of the Association for Computa-Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., tional Linguistics, 8:726–742. [find](https://scholar.google.com/scholar?q=Chanan%2C+G%2E%2C+Killeen%2C+T%2E%2C+Lin%2C+Z%2E%2C+Gimelshein%2C+N%2E%2C+Antiga%2C+L%2E%2C+et+al%2E+%282019%29%2E+Pytorch%3A+An+imperative+style%2C+high%2DLiu%2C+Y%2E%2C+Gu%2C+J%2E%2C+Goyal%2C+N%2E%2C+Li%2C+X%2E%2C+Edunov%2C+S%2E%2C+Ghazvinine%2Dperformance+deep+learning+lib)
34. Matena, M., Zhou, Y., Li, W., and Liu, P. J. (2019). [Exploring the limits of transfer learning with a unified text-Meftouh, K., Harrat, S., Jamoussi, S., Abbas, M., and to-text transformer](https://md.research.elamur.ai/free/en/papers/raffel-2019-exploring-limits-transfer-learning-unified/). arXiv preprint arXiv:1910.10683. [↗](https://arxiv.org/abs/1910.10683)
35. Smaili, K. (2015). [Machine translation experiments on PADIC: A parallel Arabic DIalect corpus](https://md.research.elamur.ai/free/en/papers/meftouh-2015-machine-translation-experiments-padic-parallel/). In Proceed-Reimers, N. and Gurevych, I. (2020). Making mono-ings of the 29th Pacific Asia Conference on Language, lingual sentence embeddings multilingual using knowl-Information and Computation, pages 26–34, Shanghai, edge distillation. In Proceedings of the 2020 Conference China, October. on Empirical Methods in Natural Language Processing (EMNLP), pages 4512–4525.
36. Mohamed, E., Mohit, B., and Oflazer, K. (2012). [Transforming standard arabic to colloquial arabic](https://md.research.elamur.ai/free/en/papers/mohamed-2012-transforming-standard-arabic-colloquial-arabic/). In Pro-Sajjad, H., Darwish, K., and Belinkov, Y. (2013). Trans-ceedings of the 50th Annual Meeting of the Association lating dialectal arabic to english. In Proceedings of the for Computational Linguistics (Volume 2: Short Papers), 51st Annual Meeting of the Association for Computa-pages 176–180. tional Linguistics (Volume 2: Short Papers), pages 1–6.
37. Mubarak, H. (2018). Dial2msa: A tweets corpus for con-Sajjad, H., Abdelali, A., Durrani, N., and Dalvi, F. (2020). verting dialectal arabic to modern standard arabic. In Arabench: Benchmarking dialectal arabic-english ma-OSACT 3: The 3rd Workshop on Open-Source Arabic chine translation. In Proceedings of the 28th Interna-Corpora and Processing Tools, page 49. tional Conference on Computational Linguistics, pages 5094–5107. [find](https://scholar.google.com/scholar?q=Mubarak%2C+H%2E+%282018%29%2E+Dial2msa%3A+A+tweets+corpus+for+con%2DSajjad%2C+H%2E%2C+Abdelali%2C+A%2E%2C+Durrani%2C+N%2E%2C+and+Dalvi%2C+F%2E+%282020%29%2E+verting+dialectal+arabic+to+modern+standard+arabic%2E+In+Arabench%3A+Benchmarking+dialect)
38. Nagoudi, E. M. B., Elmadany, A., and Abdul-Mageed, M. (2021). Investigating code-mixed Modern Standard Salloum, W. and Habash, N. (2011). Dialectal to standard Arabic-Egyptian to English machine translation. In Pro-arabic paraphrasing to improve arabic-english statistical ceedings of the Fifth Workshop on Computational Ap-machine translation. In Proceedings of the first work-proaches to Linguistic Code-Switching, pages 56–64, 10 shop on algorithms and resources for modelling of dialects and language varieties, pages 10–21. [find](https://scholar.google.com/scholar?q=Nagoudi%2C+E%2E+M%2E+B%2E%2C+Elmadany%2C+A%2E%2C+and+Abdul%2DMageed%2C+M%2E+%282021%29%2E+Investigating+code%2Dmixed+Modern+Standard+Salloum%2C+W%2E+and+Habash%2C+N%2E+%282011%29%2E+Dialectal+to+standard+Arabic%2DEgyptian+to+English+machine+trans)
39. Salloum, W. and Habash, N. (2013). [Dialectal arabic to english machine translation: Pivoting through modern standard arabic](https://md.research.elamur.ai/free/en/papers/salloum-2013-dialectal-arabic-english-machine-translation/). In Proceedings of the 2013 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 348–358.
40. Salloum, W., Elfardy, H., Alamir-Salloum, L., Habash, N., and Diab, M. (2014). Sentence level dialect identification for machine translation system selection. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 772–778. [find](https://scholar.google.com/scholar?q=Salloum%2C+W%2E%2C+Elfardy%2C+H%2E%2C+Alamir%2DSalloum%2C+L%2E%2C+Habash%2C+N%2E%2C+and+Diab%2C+M%2E+%282014%29%2E+Sentence+level+dialect+identification+for+machine+translation+system+selection%2E+In+Proceedings+of+the+52nd+Annual+Meeting)
41. Song, K., Tan, X., Qin, T., Lu, J., and Liu, T.-Y. (2019). [Mass: Masked sequence to sequence pre-training for language generation](https://md.research.elamur.ai/free/en/papers/song-2019-mass-masked-sequence-sequence-pre/). arXiv preprint arXiv:1905.02450. [↗](https://arxiv.org/abs/1905.02450)
42. Steinberger, R., Pouliquen, B., Widiger, A., Ignat, C., Erjavec, T., Tufiş, D., and Varga, D. (2006). [The jrc-acquis: A multilingual aligned parallel corpus with 20+ languages](https://md.research.elamur.ai/free/en/papers/steinberger-2006-jrc-acquis-multilingual-aligned-parallel/). In Proceedings of the Fifth International Conference on Language Resources and Evaluation (LREC’06).
43. Takezawa, T., Kikui, G., Mizushima, M., and Sumita, E. (2007). Multilingual spoken language corpus development for communication research. In International Journal of Computational Linguistics & Chinese Language Processing, Volume 12, Number 3, September 2007: Special Issue on Invited Papers from ISCSLP 2006, pages 303–324. [find](https://scholar.google.com/scholar?q=Takezawa%2C+T%2E%2C+Kikui%2C+G%2E%2C+Mizushima%2C+M%2E%2C+and+Sumita%2C+E%2E+%282007%29%2E+Multilingual+spoken+language+corpus+development+for+communication+research%2E+In+International+Journal+of+Computational+Linguistics+%26+Chine)
44. Tiedemann, J. (2012). Parallel data, tools and interfaces in opus. 2012:2214–2218. [find](https://scholar.google.com/scholar?q=Tiedemann%2C+J%2E+%282012%29%2E+Parallel+data%2C+tools+and+interfaces+in+opus%2E+2012%3A2214%E2%80%932218%2E)
45. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., and Polosukhin, I. (2017). [Attention is all you need](https://md.research.elamur.ai/free/en/papers/vaswani-2017-attention-all-need/). In Advances in neural information processing systems, pages 5998–6008.
46. Xue, L., Constant, N., Roberts, A., Kale, M., Al-Rfou, R., Siddhant, A., Barua, A., and Raffel, C. (2020). [mt5: A massively multilingual pre-trained text-to-text transformer](https://md.research.elamur.ai/free/en/papers/xue-2021-mt5-massively-multilingual-pre-trained/). arXiv preprint arXiv:2010.11934. [↗](https://arxiv.org/abs/2010.11934)
47. Zbib, R., Malchiodi, E., Devlin, J., Stallard, D., Matsoukas, S., Schwartz, R., Makhoul, J., Zaidan, O., and Callison-Burch, C. (2012). [Machine translation of arabic dialects](https://md.research.elamur.ai/free/en/papers/zbib-2012-machine-translation-arabic-dialects/). In Proceedings of the 2012 conference of the north american chapter of the association for computational linguistics: Human language technologies, pages 49–59.
48. Ziemski, M., Junczys-Dowmunt, M., and Pouliquen, B. (2016). [The united nations parallel corpus v1](https://md.research.elamur.ai/free/en/papers/ziemski-2016-united-nations-parallel-corpus-v1/). 0. In Lrec.

*Updated 1 November 2026*