Elamur Research⌕
☰
home / papers / 2024 / ZAEBUC-Spoken: A Multilingual Multidialectal Arabic-English Speech Corpus

ZAEBUC-Spoken: A Multilingual Multidialectal Arabic-English Speech Corpus

Paper · 2024 · Injy Hamed, Fadhl Eryani, David Palfreyman et al.

Summary

Speech and language tools for Arabic work best on the formal standard variety. They perform poorly on regional dialects and on conversations where speakers mix languages or varieties, and evaluation data for such speech is scarce. The authors address this gap with ZAEBUC-Spoken, a spoken extension of the earlier written ZAEBUC bilingual corpus. They recorded about twelve hours of Zoom meetings in which pairs of university students role-play a workplace task: first they brainstorm, then they pitch their ideas to an unfamiliar senior Interlocutor. Each meeting follows fixed phases, and each phase has its own expected language setup. Besides accented English, the recordings feature MSA (the formal standard form of Arabic) and two regional dialects, Gulf and Egyptian, with code-switching both between Arabic and English and among the Arabic variants. The team produced manual transcriptions under purpose-built guidelines for spontaneous, mixed-language speech. They added two annotation layers: dialectness levels for the stretches where Arabic variants are mixed, and automatic tokenization, lemmas and POS tags. The resulting public corpus is a hard benchmark for ASR and a basis for studying how bilingual speakers interact.

Key points

  • The corpus contains 14 meetings, split evenly between Arabic-speaking and English-speaking Interlocutors, plus a 15th pilot recording that covers only the student-discussion phase. In total it amounts to roughly 12 hours.
  • 16 Emirati students took part, all but one female. Each joined at most two meetings, one with each Interlocutor type. Speakers come from six nationalities overall.
  • Each meeting has four phases:
    • Moderator introduction, in English.
    • Free-language student discussion, mostly Gulf Arabic mixed with English.
    • Presentation to the Interlocutor, in either English or MSA. The Arabic sessions mix MSA with Gulf or Egyptian dialect.
    • Moderator closing, in English.
  • Every meeting has a mixed audio track. For 11 of the 15 recordings there are also separate per-speaker channels, which lets researchers handle overlapping speech.
  • Annotation layers are manual transcripts, dialectness levels for the Arabic-variant mixing portion, and automatic morphology (tokenization, lemmatization, POS). The corpus is released at zaebuc.org.

Interesting ideas

  • I1. The data was elicited through a staged scenario in which each phase has a planned language setup, so the kind of code-switching expected in each segment is known in advance, which turns metadata into a free stratification for analysis.
  • I2. Instead of a binary MSA-versus-dialect label, mixed stretches get a graded dialectness level, which captures that speech often sits between varieties.
  • I3. The authors note that tagging spontaneous speech is harder than tagging text, because repetitions and self-corrections interrupt normal sentence flow, and they say their transcription guidelines handle such disfluencies explicitly.
  • I4. The corpus keeps both a mixed recording and per-speaker channels, so different downstream uses (overlap handling vs. clean single-speaker input) can share one resource.
  • I5. The authors release automatic annotations openly labelled as unrevised, with manual correction planned later, so the resource is usable now and its quality status stays transparent.

For Mova

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

  • I1 (learn): Evaluation sets could be built from controlled scenarios whose segments differ in register (formal vs. chatty English), so scores can be broken down by register instead of averaged over everything.
  • I2 (tokens): A graded "informality/dialect" score per sentence could warn the tagger and lemmatizer when input drifts from standard written English, so explanations can flag less trustworthy analyses.
  • I3 (tree): Transcribed or chat-like inputs could get an explicit disfluency convention (marking restarts and repairs), so the UD parser attaches reparanda predictably instead of producing broken trees.
  • I4 (world): For multi-party dialogues, keeping utterances tagged per speaker alongside the merged transcript would let the world model track who holds which belief or proposal.
  • I5 (learn): Silver data from witness parsers could carry an explicit "automatic, unrevised" provenance tag per layer, so training and negative-control evaluations never confuse silver with gold.

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, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 4Research subjects and languages the work is about.English · 2,263 worksSemitic languages · 2,908 worksMultilingual & cross-lingual NLP · 828 worksSpeech & spoken language · 480 works

Cite

Cite as Injy Hamed, Fadhl Eryani, David Palfreyman and Nizar Habash. 2024. ZAEBUC-Spoken: A Multilingual Multidialectal Arabic-English Speech Corpus. Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024).

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

@inproceedings{hamed2024zaebuc,
  title = {{ZAEBUC-Spoken: A Multilingual Multidialectal Arabic-English Speech Corpus}},
  author = {Injy Hamed, Fadhl Eryani, David Palfreyman and Nizar Habash},
  booktitle = {Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)},
  year = {2024},
  doi = {10.63317/29jtxbczvc9i},
  url = {https://doi.org/10.63317/29jtxbczvc9i}
}

References

54 references, 20 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Faisal Al-Shargi, Aidan Kaplan, Ramy Eskander, Nizar Habash, and Owen Rambow. 2016. Morphologically Annotated Corpus and a Morphological Analyzer for Moroccan and San’ani Yemeni Arabic. In Proceedings of LREC, pages 1300– 1306. find
  2. Ahmed Ali, Peter Bell, James Glass, Yacine Messaoui, Hamdy Mubarak, Steve Renals, and Yifan Zhang. 2016. The MGB-2 challenge: Arabic multi-dialect broadcast media recognition. In SLT, pages 279–284.
  3. Ahmed Ali, Suwon Shon, Younes Samih, Hamdy Mubarak, Ahmed Abdelali, James Glass, Steve Renals, and Khalid Choukri. 2019. The MGB-5 challenge: Recognition and dialect identification of dialectal Arabic speech. In Proceedings of ASRU, pages 1026–1033. find
  4. Ahmed Ali, Stephan Vogel, and Steve Renals. 2017. Speech recognition challenge in the wild: Arabic MGB-3. In Proceedings of the Automatic Speech Recognition and Understanding Workshop, pages 316–322.
  5. Khalid Almeman, Mark Lee, and Ali Abdulrahman Almiman. 2013. Multi dialect Arabic speech parallel corpora. In Proceedings of the International Conference on Communications, Signal Processing, and their Applications (ICCSPA), pages 1–6. find
  6. Djegdjiga Amazouz, Martine Adda-Decker, and Lori Lamel. 2018. The French-Algerian code-switching triggered audio corpus (FACST). In Proceedings of LREC, pages 1468–1473. find
  7. Mohamed Balabel, Injy Hamed, Slim Abdennadher, Ngoc Thang Vu, and Özlem Çetinoğlu. 2020. Cairo student code-switch (CSCS) corpus: An annotated Egyptian Arabic-English corpus. In Proceedings of LREC, pages 3973–3977.
  8. Kalika Bali, Jatin Sharma, Monojit Choudhury, and Yogarshi Vyas. 2014. I am borrowing ya mixing?" an analysis of English-Hindi code mixing in facebook. In Proceedings of the first workshop on computational approaches to code switching, pages 116–126.
  9. Houda Bouamor, Nizar Habash, Mohammad Salameh, Wajdi Zaghouani, Owen Rambow, Dana Abdulrahim, Ossama Obeid, Salam Khalifa, Fadhl Eryani, Alexander Erdmann, and Kemal Oflazer. 2018. The MADAR Arabic Dialect Corpus and Lexicon. In Proceedings of LREC, pages 3387–3396. find
  10. Hennie Brugman, Albert Russel, and Xd Nijmegen. 2004. Annotating multi-media/multi-modal resources with ELAN. In Proceedings of LREC, pages 2065–2068. find
  11. Shammur A Chowdhury, Younes Samih, Mohamed Eldesouki, and Ahmed Ali. 2020. Effects of dialectal code-switching on speech modules: A study using Egyptian Arabic broadcast speech. In Proceedings of Interspeech, pages 2382– 2386. find
  12. Shammur Absar Chowdhury, Amir Hussein, Ahmed Abdelali, and Ahmed Ali. 2021. Towards one model to rule all: Multilingual strategy for dialectal code-switching Arabic ASR. In Proceedings of Interspeech, pages 2466–2470.
  13. A Seza Doğruöz, Sunayana Sitaram, Barbara E Bullock, and Almeida Jacqueline Toribio. 2021. A survey of code-switching: Linguistic and social perspectives for language technologies. In Proceedings of ACL-IJCNLP, pages 1654–1666.
  14. Mohamed Elmahdy, Mark Hasegawa-Johnson, and Eiman Mustafawi. 2014. Development of a TV broadcasts speech recognition system for Qatari Arabic. In Proceedings of LREC, pages 3057– 3061. word, Charles A Ferguson. 1959. 15(2):325–340. Diglossia.
  15. Björn Gambäck and Amitava Das. 2016. Comparing the level of code-switching in corpora. In Proceedings of LREC, pages 1850–1855. find
  16. Marwa Gaser, Manuel Mager, Injy Hamed, Nizar Habash, Slim Abdennadher, and Ngoc Thang Vu. 2022. Exploring segmentation approaches for neural machine translation of code-switched Egyptian Arabic-English text. In Proceedings of EACL, pages 86–100. find
  17. Nizar Habash, Fadhl Eryani, Salam Khalifa, Owen Rambow, Dana Abdulrahim, Alexander Erdmann, Reem Faraj, Wajdi Zaghouani, Houda Bouamor, Nasser Zalmout, Sara Hassan, Faisal Al shargi, Sakhar Alkhereyf, Basma Abdulkareem, Ramy Eskander, Mohammad Salameh, and Hind Saddiki. 2018. Unified guidelines and resources for Arabic dialect orthography. In Proceedings of LREC, pages 3628–3637.
  18. Nizar Habash and David Palfreyman. 2022. ZAE-BUC: An annotated Arabic-English bilingual writer corpus. In Proceedings of LREC, pages 79–88. find
  19. Nizar Habash, Owen Rambow, Mona Diab, and Reem Kanjawi-Faraj. 2008. Guidelines for annotation of Arabic dialectness. In Proceedings of the LREC Workshop on HLT & NLP within the Arabic world, pages 49–53. find
  20. 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. ↗
  21. Nizar Y Habash. 2010. Introduction to Arabic natural language processing, volume 3. Morgan & Claypool Publishers. find
  22. Injy Hamed, Pavel Denisov, Chia-Yu Li, Mohamed Elmahdy, Slim Abdennadher, and Ngoc Thang Vu. 2021. Investigations on speech recognition systems for low-resource dialectal Arabic– English code-switching speech. Computer Speech & Language, 72:101278. ↗
  23. Injy Hamed, Mohamed Elmahdy, and Slim Abdennadher. 2018. Collection and analysis of code-switch Egyptian Arabic-English speech corpus. In Proceedings of LREC, pages 3805–3809. find
  24. Injy Hamed, Nizar Habash, Slim Abdennadher, and Ngoc Thang Vu. 2022. ArzEn-ST: A three-way speech translation corpus for code-switched Egyptian Arabic-English. In Proceedings of the Arabic Natural Language Processing Workshop (WANLP), pages 119–130. ↗
  25. Injy Hamed, Ngoc Thang Vu, and Slim Abdennadher. 2020. ArzEn: A speech corpus for code-switched Egyptian Arabic-English. In Proceedings of LREC, pages 4237–4246. find
  26. Go Inoue, Salam Khalifa, and Nizar Habash. 2021. Morphosyntactic tagging with pre-trained language models for Arabic and its dialects. In Findings of ACL, pages 1708—-1719.
  27. Manal A Ismail. 2015. The sociolinguistic dimensions of code-switching between Arabic and English by Saudis. International Journal of English Linguistics, 5(5):99. ↗
  28. 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. ↗
  29. Amr Keleg, Sharon Goldwater, and Walid Magdy. 2023. ALDi: Quantifying the Arabic level of dialectness of text. In Proceedings of EMNLP.
  30. Salam Khalifa, Nizar Habash, Dana Abdulrahim, and Sara Hassan. 2016. A Large Scale Corpus of Gulf Arabic. In Proceedings of LREC, pages 4282–4289.
  31. Salam Khalifa, Nizar Habash, Fadhl Eryani, Ossama Obeid, Dana Abdulrahim, and Meera Al Kaabi. 2018. A morphologically annotated corpus of Emirati Arabic. In Proceedings of LREC, pages 3839–3846.
  32. Mohamed Maamouri, Ann Bies, Tim Buckwalter, Mona Diab, Nizar Habash, Owen Rambow, and Dalila Tabessi. 2006. Developing and Using a Pilot Dialectal Arabic Treebank. In Proceedings of LREC, pages 443–448.
  33. 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. find
  34. Mohamed Maamouri, Ann Bies, Seth Kulick, Michael Ciul, Nizar Habash, and Ramy Eskander. 2014. Developing an Egyptian Arabic Treebank: Impact of Dialectal Morphology on Annotation and Tool Development. In Proceedings of LREC, pages 2348–2354. find
  35. Hamdy Mubarak, Amir Hussein, Shammur Absar Chowdhury, and Ahmed Ali. 2021. QASR: QCRI Aljazeera speech resource–a large scale annotated Arabic speech corpus. In Proceedings of ACL, pages 2274–2285. Joakim Nivre, Željko Agić, Lars Ahrenberg, Maria Jesus Aranzabe, Masayuki Asahara, Aitziber Atutxa, Miguel Ballesteros, John Bauer, Kepa Bengoetxea, Riyaz Ahmad Bhat, Eckhard Bick, Cristina Bosco, Gosse Bouma, Sam Bowman, Marie Candito, Gülşen Cebiroğlu Eryiğit, Giuseppe G. A. Celano, Fabricio Chalub, Jinho Choi, Çağrı Çöltekin, Miriam Connor, Elizabeth Davidson, Marie-Catherine de Marneffe, Valeria de Paiva, Arantza Diaz de Ilarraza, Kaja Dobrovoljc, Timothy Dozat, Kira Droganova, Puneet Dwivedi, Marhaba Eli, Tomaž Erjavec, Richárd Farkas, Jennifer Foster, Cláudia Freitas, Katarína Gajdošová, Daniel Galbraith, Marcos Garcia, Filip Ginter, Iakes Goenaga, Koldo Gojenola, Memduh Gökırmak, Yoav Goldberg, Xavier Gómez Guinovart, Berta Gonzáles Saavedra, Matias Grioni, Normunds Grūzı̄tis, Bruno Guillaume, Nizar Habash, Jan Hajič, Linh Hà Mȳ, Dag Haug, Barbora Hladká, Petter Hohle, Radu Ion, Elena Irimia, Anders Johannsen, Fredrik Jørgensen, Hüner Kaşıkara, Hiroshi Kanayama, Jenna Kanerva, Natalia Kotsyba, Simon Krek, Veronika Laippala, Phuong Lê Hồng, Alessandro Lenci, Nikola Ljubešić, Olga Lyashevskaya, Teresa Lynn, Aibek Makazhanov, Christopher Manning, Cătălina Mărănduc, David Mareček, Héctor Martínez Alonso, André Martins, Jan Mašek, Yuji Matsumoto, Ryan McDonald, Anna Missilä, Verginica Mititelu, Yusuke Miyao, Simonetta Montemagni, Amir More, Shunsuke Mori, Bohdan Moskalevskyi, Kadri Muischnek, Nina Mustafina, Kaili Müürisep, Luong Nguyễn Thi., Huyền Nguyễn Thi. Minh, Vitaly Nikolaev, Hanna Nurmi, Stina Ojala, Petya Osenova, Lilja Øvrelid, Elena Pascual, Marco Passarotti, Cenel-Augusto Perez, Guy Perrier, Slav Petrov, Jussi Piitulainen, Barbara Plank, Martin Popel, Lauma Pretkalnin , a, Prokopis Prokopidis, Tiina Puolakainen, Sampo Pyysalo, Alexandre Rademaker, Loganathan Ramasamy, Livy Real, Laura Rituma, Rudolf Rosa, Shadi Saleh, Manuela Sanguinetti, Baiba Saulı̄te, Sebastian Schuster, Djamé Seddah, Wolfgang Seeker, Mojgan Seraji, Lena Shakurova, Mo Shen, Dmitry Sichinava, Natalia Silveira, Maria Simi, Radu Simionescu, Katalin Simkó, Mária Šimková, Kiril Simov, Aaron Smith, Alane Suhr, Umut Sulubacak, Zsolt Szántó, Dima Taji, Takaaki Tanaka, Reut Tsarfaty, Francis Tyers, Sumire Uematsu, Larraitz Uria, Gertjan van Noord, Viktor Varga, Veronika Vincze, Jonathan North Washington, Zdeněk Žabokrtský, Amir Zeldes, Daniel Zeman, and Hanzhi Zhu. 2017. Universal dependencies 2.0. ↗
  36. 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 LREC, pages 7022– 7032.
  37. David Palfreyman and Nizar Y Habash. 2022. Bilingual writers and corpus analysis. Taylor & Francis Group. ↗
  38. Daniel Povey, Arnab Ghoshal, Gilles Boulianne, Lukás̆ Burget, Ondr̆ej Glembek, Nagendra Goel, Mirko Hannemann, Petr Motlíc̆ek, Yanmin Qian, Petr Schwarz, Jan Silovský, Georg Stemmer, and Karel Veselý. 2011. The Kaldi speech recognition toolkit. In Proceedings of ASRU, pages 1–4. find
  39. Adithya Pratapa, Gayatri Bhat, Monojit Choudhury, Sunayana Sitaram, Sandipan Dandapat, and Kalika Bali. 2018. Language modeling for code-mixing: The role of linguistic theory based synthetic data. In Proceedings of ACL, pages 1543– 1553.
  40. 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 ACL: System Demonstration, pages 101–108.
  41. Surangika Ranathunga and Nisansa de Silva. 2022. Some languages are more equal than others: Probing deeper into the linguistic disparity in the NLP world. In Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing, pages 823–848. ↗
  42. Colleen Richey, Cynthia D’Angelo, Nonye Alozie, Harry Bratt, and Elizabeth Shriberg. 2016. The SRI speech-based collaborative learning corpus. In Proceedings of Interspeech, pages 1550– 1554. find
  43. Dima Taji, Nizar Habash, and Daniel Zeman. 2017. Universal dependencies for Arabic. In Proceedings of the Workshop for Arabic Natural Language Processing (WANLP), pages 166–176.
  44. Omar Zaidan and Chris Callison-Burch. 2011. The Arabic online commentary dataset: an annotated dataset of informal Arabic with high dialectal content. In Proceedings of the Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, pages 37–41. 10.
  45. Language Resource References Pty Ltd, Sydney, and Australia Appen. 2006a. Gulf Arabic conversational telephone speech LDC2006S43. Web Download. Philadelphia: Linguistic Data Consortium. find
  46. Pty Ltd, Sydney, and Australia Appen. 2006b. Gulf Arabic conversational telephone speech, transcripts LDC2006T15. Web Download. Philadelphia: Linguistic Data Consortium. find
  47. Pty Ltd, Sydney, and Australia Appen. 2007. Levantine Arabic conversational telephone speech, transcripts LDC2007T01. Web Download. Philadelphia: Linguistic Data Consortium. find
  48. BBN Technologies, John Makhoul, Bushra Zawaydeh, Frederick Choi, and David Stallard. 2005. BBN/AUB DARPA Babylon Levantine Arabic speech and transcripts LDC2005S08. Web Download. Philadelphia: Linguistic Data Consortium. find
  49. Hassan Gadalla, Hanaa Kilany, Howaida Arram, Ashraf Yacoub, Alaa El-Habashi, Amr Shalaby, Krisjanis Karins, Everett Rowson, Robert MacIntyre, Paul Kingsbury, David Graff, and Cynthia McLemore. 1997. CALLHOME Egyptian Arabic transcripts. Linguistic Data Consortium, Philadelphia. find
  50. H. Kilany, H. Gadalla, H. Arram, A. Yacoub, A. El-Habashi, and C. McLemore. 2002. Egyptian Colloquial Arabic Lexicon. LDC catalog number LDC99L22. find
  51. Mohamed Maamouri, Ann Bies, Seth Kulick, Dalila Tabessi, and Sondos Krouna. 2012. Egyptian Arabic Treebank DF Parts 1-8 V2.0 - LDC catalog numbers LDC2012E93, LDC2012E98, LDC2012E89, LDC2012E99, LDC2012E107, LDC2012E125, LDC2013E12, LDC2013E21. find
  52. Mohamed Maamouri, Tim Buckwalter, Dave Graff, and Hubert Jin. 2007. Fisher Levantine Arabic conversational telephone speech LDC2007S02. Web Download. Philadelphia: Linguistic Data Consortium. find
  53. Paul Boersma and David Weenink. 2022. Praat: doing phonetics by computer [Computer program]. Version 6.2.14, retrieved from https://www.praat.org. ↗
  54. Colleen Richey, Nonye Alozie Cynthia D’Angelo, Harry Bratt, and Elizabeth Shriberg. 2019. SRI speech-based collaborative learning corpus LDC2019S01. Web Download. Philadelphia: Linguistic Data Consortium. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/hamed-2024-zaebuc-spoken-multilingual-multidialectal-arabic/