Elamur Research⌕
☰
home / papers / 2020 / Cairo Student Code-Switch (CSCS) Corpus: An Annotated Egyptian Arabic-English Corpus

Cairo Student Code-Switch (CSCS) Corpus: An Annotated Egyptian Arabic-English Corpus

Summary

Code-switching (CS) between languages is common in Arabic-speaking communities, yet NLP tools for it remain weak, largely because annotated data is scarce. This is especially true for dialectal Arabic, whose spelling and grammar have no fixed standard. The authors build CSCS, a corpus of Egyptian Arabic-English speech transcriptions. It draws on interviews from Hamed et al. (2018), using recordings of six of the original twelve participants.

The work had two phases:

  • Transcription: the authors wrote orthographic guidelines, and three transcribers followed them.
  • Gold annotation: two native-Arabic authors added tokenization, UD POS tags and lemmas in CoNLL-U format, then checked each other's work.

The main output has two parts. One is the annotated corpus, which the authors describe as the first Arabic CS resource covering all three layers. The other is a public set of guidelines for the specific problems of annotating mixed-language data. The resource adds a language pair not previously represented among morpho-syntactically annotated CS corpora. It also lays the groundwork for a future Egyptian Arabic-English UD treebank.

Key points

  • The source speech comes from 12 interviewees at the German University in Cairo, totalling 5.3 hours. CSCS uses the data of six of these speakers.
  • Annotation layers are tokenization, UD POS tagging and lemmatization, done in that order and stored in CoNLL-U. An earlier pilot had tagged 388 sentences, but without proper guidelines and without the UD tagset.
  • The spelling rules build on the conventions of the Egyptian Arabic Wikipedia community rather than Habash et al. (2012). Where Wikipedia leaves options open, the authors usually narrow the choice to a single variant.
  • The annotation guidelines are organised around three operations: unification, discrimination and disambiguation.
  • The data mixes Egyptian Arabic and English only; no MSA-dialect switching was observed. Dependency relations are planned as future work, which would turn the corpus into a treebank.

Interesting ideas

  • I1. Placing two languages side by side in one sentence exposes places where their separately built UD treebanks treat the same structure differently; mixed data thus works as a consistency probe. — For Mova (tree): when converting or merging UD dialects and treebanks, gather sentences where the same construction appears under both conventions, and flag disagreements as candidate conversion rules.
  • I2. The authors chose their spelling standard to match the biggest freely available text source (Wikipedia), not the most academically established one, so the corpus can be combined with web data. — For Mova (tokens): normalise variant spellings toward the convention of whatever large corpus feeds the lexicon and lemma tables, so lookups hit more often.
  • I3. Where conventions allow several spellings, the authors fix one variant per case to cut ambiguity at the source rather than handling it later. — For Mova (tokens): keep a canonical-form table so the tagger and lemmatizer see one variant per word, which shrinks sparse counts and makes errors easier to trace.
  • I4. Each annotation decision is framed as one of a few named operations (unify, discriminate, disambiguate), giving a reusable checklist for cross-language cases. — For Mova (learn): tag each self-induced repair rule with the kind of operation it performs, so rule sets can be audited and conflicting rules spotted.
  • I5. Annotators work in a fixed order (tokens, then tags, then lemmas), each one revises the other's share, and disagreements are discussed until consensus. — For Mova (learn): run silver-data checking in the same staged way, with witnesses cross-reviewing each layer before the next is built, and log the resolved conflicts as gold examples.

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: Universal Dependencies · Treebanks · Part-of-speech tagging

Cited by

6 papers of our library cite it (of the 140884 whose bibliographies we read); 1 of them have a page here, 5 not yet.

Cite

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

@inproceedings{balabel2020cairo,
  title = {{Cairo Student Code-Switch (CSCS) Corpus: An Annotated Egyptian Arabic-English Corpus}},
  author = {Balabel, Mohamed and Hamed, Injy and Abdennadher, Slim and Vu, Ngoc Thang and Çetinoğlu, Özlem},
  booktitle = {Proceedings of the Twelfth Language Resources and Evaluation Conference},
  year = {2020},
  url = {https://aclanthology.org/2020.lrec-1.489/}
}

References

25 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Abainia, K. (2019). Dzdc12: a new multipurpose parallel algerian arabizi–french code-switched corpus. Language Resources and Evaluation, pages 1–37. find
  2. Amazouz, D., Adda-Decker, M., and Lamel, L. (2018). The french-algerian code-switching triggered audio corpus (facst). find
  3. Bhat, I., Bhat, R. A., Shrivastava, M., and Sharma, D. (2017). Joining hands: Exploiting monolingual treebanks for parsing of code-mixing data. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers, volume 2, pages 324–330. find
  4. Bhat, I., Bhat, R. A., Shrivastava, M., and Sharma, D. (2018). Universal dependency parsing for hindi-english code-switching. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), volume 1, pages 987–998. find
  5. Çetinoğlu, Ö. (2016). A turkish-german code-switching corpus. In Nicoletta Calzolari (Conference Chair), et al., editors, Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC 2016), Paris, France, may. European Language Resources Association (ELRA). find
  6. Çetinoğlu, Ö. and Çöltekin, Ç. (2016). Part of speech annotation of a Turkish-German code-switching corpus. In Proceedings of the 10th Linguistic Annotation Workshop held in conjunction with ACL 2016 (LAW-X 2016), pages 120–130, Berlin, Germany, August. Association for Computational Linguistics. find
  7. Çetinoğlu, Ö. and Çöltekin, Ç. (2019). Challenges of annotating a code-switching treebank. In Proceedings of the 18th International Workshop on Treebanks and Linguistic Theories (TLT, SyntaxFest 2019), pages 82–90. find
  8. Çetinoğlu, Ö., Schulz, S., and Vu, N. T. (2016). Challenges of computational processing of code-switching. In Proceedings of the Second Workshop on Computational Approaches to Code Switching, pages 1–11. find
  9. Çetinoğlu, Ö. (2017). A code-switching corpus of Turkish-German conversations. In Proceedings of the 11th Linguistic Annotation Workshop, pages 34–40, Valencia, Spain, April. Association for Computational Linguistics. find
  10. Cotterell, R., Renduchintala, A., Saphra, N., and Callison-Burch, C. (2014). An algerian arabic-french code-switched corpus. In Workshop on Free/Open-Source Arabic Corpora and Corpora Processing Tools Workshop Programme, page 34. find
  11. Elfardy, H. and Diab, M. (2012). Token level identification of linguistic code switching. In Proceedings of COLING 2012: Posters, pages 287–296. find
  12. Habash, N., Diab, M. T., and Rambow, O. (2012). Conventional orthography for dialectal arabic. In LREC, pages 711–718. find our card
  13. Hajic, J., Smrz, O., Zemánek, P., Šnaidauf, J., and Beška, E. (2004). Prague arabic dependency treebank: Development in data and tools. In Proc. of the NEMLAR Intern. Conf. on Arabic Language Resources and Tools, pages 110–117. find
  14. Hamed, I., Elmahdy, M., and Abdennadher, S. (2017). Building a first language model for code-switch arabic-english. Procedia Computer Science, 117:208–216. find
  15. Hamed, I., Elmahdy, M., and Abdennadher, S. (2018). Collection and analysis of code-switch egyptian arabic-english speech corpus. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC-2018). find
  16. Ismail, M. A. (2015). The sociolinguistic dimensions of code-switching between arabic and english by saudis. International Journal of English Linguistics, 5(5):99. find
  17. Mohdeb-Amazouz, D., Martine, A.-D., and Lamel, L. (2016). Arabic-french code-switching across maghreb arabic dialects: a quantitative analysis. find
  18. Mustafa, M. and Suleman, H. (2011). Building a multilingual and mixed arabic-english corpus. In Proceedings Arabic Language Technology International Conference (ALTIC). find
  19. Nivre, J., De Marneffe, M.-C., Ginter, F., Goldberg, Y., Hajic, J., Manning, C. D., McDonald, R., Petrov, S., Pyysalo, S., Silveira, N., et al. (2016). Universal dependencies v1: A multilingual treebank collection. In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC’16), pages 1659–1666. find our card
  20. Plank, B., Hovy, D., and Søgaard, A. (2014). Learning part-of-speech taggers with inter-annotator agreement loss. In Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, pages 742–751. find our card
  21. Poplack, S. (1980). Sometimes iâll start a sentence in spanish y termino en espanol: toward a typology of code-switching1. Linguistics, 18(7-8):581–618. find
  22. Sabty, C., Elmahdy, M., and Abdennadher, S. (2019). Named entity recognition on arabic-english code-mixed data. In 2019 IEEE 13th International Conference on Semantic Computing (ICSC), pages 93–97. IEEE. find
  23. Samih, Y. and Maier, W. (2016). An arabic-moroccan darija code-switched corpus. In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC 2016), pages 4170–4175. find
  24. Solorio, T. and Liu, Y. (2008). Part-of-speech tagging for english-spanish code-switched text. In Proceedings of the Conference on Empirical Methods in Natural Language Processing, pages 1051–1060. Association for Computational Linguistics. find
  25. Taji, D., Habash, N., and Zeman, D. (2017). Universal dependencies for arabic. In Proceedings of the Third Arabic Natural Language Processing Workshop, pages 166– 176. find our card

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/balabel-2020-cairo-student-code-switch-cscs/