Elamur Research⌕
☰
home / papers / 2012 / Open-Source Boundary-Annotated Corpus for Arabic Speech and Language Processing

Open-Source Boundary-Annotated Corpus for Arabic Speech and Language Processing

Paper · 2012 · Claire Brierley, Majdi Sawalha and Eric Atwell

Summary

The paper addresses a gap in Arabic text-to-speech research. Phrase break prediction, which decides where a speaker would naturally chunk text, needs training data that is both part-of-speech tagged and boundary-annotated, and no such resource existed for Arabic. English speech corpora mark perceived breaks after the fact. The authors instead take the prescriptive pause system of Tajwīd recitation in the Qur'an (ḥafṣ bin ʿāṣim style) and treat it as prosodic annotation. They combine stop marks from the Tanzil project with PoS tags from the Qur'anic Arabic Corpus (QAC). Eight boundary strengths are reduced to the SEC-style set {major, minor, none} and also to a binary break/non-break label. The text is split into sentences at compulsory and recommended stops. The result, version 1.0 of the Boundary-Annotated Qur'an, has 77,430 words and 8230 sentences in a tab-separated, open-source format. A smaller Modern Standard Arabic corpus is planned for testing transfer. Together these resources let researchers study how prosody and syntax interact in Arabic.

Key points

  • The corpus has 77,430 words in 8230 sentences, built by merging Tanzil boundary stops with tags from an earlier QAC release (Dukes, 2010).
  • Tajwīd marks cover three major types, four minor types and one prohibited stop. These are collapsed into {major, minor, none} and further into {breaks, non-breaks}.
  • Syntax is given at two levels: the traditional Arabic {noun, verb, particle} tags and the QAC tagset with 10 major categories.
  • Each word occupies one row in a column file. The columns hold sura and aya numbers, Othmani and MSA script, both PoS levels, the raw Tajwīd symbol, both boundary labels, a sentence-terminal flag and a word-level English gloss.
  • Evaluation uses Balanced Classification Rate, because non-breaks dominate the data. The MSA test text comes from the Corpus of Contemporary Arabic, was automatically vowelised by the SALMA Tagger, and was split into sentences by hand.

Interesting ideas

  • I1. Reuse an existing prescriptive notation tradition (here, recitation pause marks) as ready-made annotation instead of paying humans to label perceived breaks.
  • I2. Keep both the fine-grained original label and its collapsed versions as parallel columns in one row, so experiments can switch granularity without rebuilding the data.
  • I3. Score a task dominated by one class with a metric that weights true positives and true negatives equally (Balanced Classification Rate), not raw accuracy.
  • I4. When punctuation is sparse, define sentence units from the strongest boundary marks the text does provide.
  • I5. The authors check that source chunk boundaries reappear as punctuation in a respected translation, which shows the chunks carry meaning across languages.

For Mova

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

  • I1 (learn): Existing markup such as editorial punctuation or style-guide comma rules could act as an extra witness when building silver chunk or clause-boundary data.
  • I2 (tree): Treebank conversion could keep the source-dialect relation next to the UD label and a coarse class, which keeps every mapping traceable.
  • I3 (learn): Evaluating rare decisions like clause breaks, irony flags or coreference links with balanced rates would stop a "predict nothing" baseline from looking strong.
  • I4 (tokens): The sentence splitter could fall back to strong-boundary cues such as colons, dashes or discourse connectives in long run-on texts like fables, so parses stay manageable.
  • I5 (back): Generation from the meaning graph could place commas and sentence breaks at frame boundaries, and the compressor's output could be checked against the original punctuation.

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 LeedsLeeds · United Kingdom40 works · 1,822 citations🎓University of JordanAmman · Jordan14 works · 171 citations

 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 1Research subjects and languages the work is about.Semitic languages · 2,908 works

Cite

Cite as Claire Brierley, Majdi Sawalha and Eric Atwell. 2012. Open-Source Boundary-Annotated Corpus for Arabic Speech and Language Processing. Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12).

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

@inproceedings{brierley2012open,
  title = {{Open-Source Boundary-Annotated Corpus for Arabic Speech and Language Processing}},
  author = {Claire Brierley, Majdi Sawalha and Eric Atwell},
  booktitle = {Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)},
  year = {2012},
  doi = {10.63317/3rc2o8rpyfcb},
  url = {https://doi.org/10.63317/3rc2o8rpyfcb}
}

References

34 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Ali, A.Y. 2000. The Holy Quran (translated by Abdullah Yusuf Ali). Hertfordshire. Wordsworth Editions Ltd. find
  2. Al-'ashmuni, Ahmad bin Muhammad Abdul-Kareem. 1973. ام صيخلتل دصقملا هعمو ،ادتبلإاو فقىلا نايب يف يدهلا رانم ءادتبلااو فقىلا يف دشرملا يف manar al-huda fi bayan al-waqf wa al-'ibtida' Mustafa Al-baabi Al-halabi. find
  3. Al-Ghalayyni. 2005. تيبرعلا سوردلا عماج "Jami' Al-Duroos Al-Arabia" Saida - Lebanon: Al-Maktaba Al-Asriyiah " تيرصعلا تبتكملا ". find
  4. Al-Sulaiti, L., Atwell, E. 2006. „The design of a corpus of contemporary Arabic.‟ In International Journal of Corpus Linguistics. Vol. 11, pp. 135-171. find
  5. Ananthakrishnan, S. and Narayanan, S.S. 2008. „Automatic Prosodic Event Detection Using Acoustic, Lexical, and Syntactic Evidence.‟ IEEE Transactions on Audio, Speech, and Language Processing, TASLP 2008. Vol. 16.1: 216-228. find
  6. Atwell, E., Brierley, C., Dukes, K., Sawalha, M. and Sharaf, A.M. 2011. „An Artificial Intelligence Approach to Arabic and Islamic Content on the Internet.‟ National Information Technology Symposium (NITS). Riyadh, Saudi Arabia. find
  7. Beckman, M. and Hirschberg, J. 1994. The ToBI annotation conventions. The Ohio State University and AT&T Bell Laboratories, unpublished manuscript. Online. Accessed September ftp://ftp.ling.ohio-state.edu/pub/phonetics/TOBI/ToBI/ToBI.6.html. find
  8. Bird, S., Klein, E. and Loper, E. 2009. Natural Language Processing with Python. Sebastopol, CA. O‟Reilly Media, Inc. find
  9. Brierley, C. 2011. Prosody Resources and Symbolic Prosodic Feartures for Automated Phrase Break Prediction. PhD Thesis. School of Computing. University of Leeds. find
  10. Brierley, C. and Atwell, E. 2010. „ProPOSEC: a Prosody and PoS Spoken English Corpus.‟ In Proceedings of LREC'10: Language Resources and Evaluation Conference, Valetta, Malta. May 2010. find
  11. Brierley, C. and Atwell, E. „ProPOSEL: A prosody and POS English lexicon for language engineering.‟ In Proceedings of LREC'08: Language Resources and Evaluation Conference, Marrakech, Morocco. May 2008.
  12. Croft, W. 1995. „Intonation Units and Grammatical Structure.‟ Linguistics. 33: 839-882 find
  13. Denny, F.M. 1976. Review [untitled]. Journal for the Scientific Study of Religion. 15.3: 287-289 find
  14. Denny, F.M. 1989. „Qur‟an Recitation: A Tradition of Oral Performance and Transmission.‟ Oral Tradition. 4/1-2: 5-26 find
  15. Dukes, K. 2011. The Qur’anic Arabic Corpus. Online. Accessed: August 2011. http://corpus.quran.com ↗
  16. Dukes, K. and Habash, N. 2010. „Morphological Annotation of Qur‟anic Arabic.‟ In Proceedings of Language Resources and Evaluation Conference (LREC 2010), Valletta, Malta. find
  17. Gilchrist, J. 2011. „Jam‟ Al-Qur‟an: The Codification of the Qur‟an Text.‟ Online. Accessed September 2011. http://www.answering-islam.org/Gilchrist/Jam/index.html ↗
  18. Grabe, E. 2001. „Prosodic Annotation.‟ PowerPoint. 9th ELSNET European Summer School on Language and Speech Communication, Prague. Accessed: 2006. find
  19. Islamic Bulletin. 2012. The Holy Quran Color Coded with Tajweed Rules. Online. Accessed: Feb. 2012. http://www.islamicbulletin.com/services/details.aspx?id=260 ↗
  20. Ladd, R. 1996. Intonational Phonology Cambridge. Cambridge University Press. find
  21. Liberman, M.Y. and Church, K.W. 1992. „Text Analysis and Word Pronunciation in Text-to-Speech Synthesis.‟ In Advances in Speech Signal Processing. Furui S. and Sondhi, M.M. (eds.). New York. Marcel Dekker Inc. find
  22. Maamouri, M., Bies, A., Buckwalter, T. and Mekki, W. 2004. The Penn Arabic Treebank: Building a Large-Scale Annotated Corpus. Philadelphia. Linguistic Data Consortium. find
  23. Ostendorf, M., Price, P. and Shattuck-Hufnagel, S. 1996. Boston University Radio Speech Corpus. Philadelphia. Linguistic Data Consortium. find
  24. Roach, P. 2000. English Phonetics and Phonology: A Practical Course (3rd. edition). Cambridge. find
  25. Cambridge University Press Ryding, Karin C. 2005. A Reference Grammar of Modern Standard Arabic. Cambridge. ↗
  26. Cambridge University Press. 2011. find
  27. Sawalha, M., Brierley, C., and Atwell, E. 2012. „Predicting Phrase Breaks in Classical and Modern Standard Arabic Text.‟ In Proceedings of LREC 2012: Language Resources and Evaluation Conference. Istanbul, Turkey. May 2012.
  28. Sawalha, M. 2011a. The SALMA – Gold Standard. Accessed: September 2011 http://www.comp.leeds.ac.uk/sawalha/goldstandard.html ↗
  29. Sawalha, Majdi. 2011b. Open-source Resources and Standards for Arabic Word Structure Analysis. Leeds: University of Leeds PhD. find
  30. Sawalha, M. and Atwell, E. 2010. „Fine-Grain Morphological Analyzer and Part-of-Speech Tagger for Arabic Text.‟ In Proceedings of LREC'10: Language Resources and Evaluation Conference, Valetta, Malta. May 2010.
  31. Sharaf, Jamal Ad-Deen Muhammad. 2004. تباحصلا فحصم ةردلاو تيبطاشلا قيرط هم ةرتاىتملا رشعلا ثاءارقلا يف mushaf as-sahabah fi al-qira'at al-'ashr al-mutawatirah min tariq ash-shatibyyah wa al-durrah Tanta: Dar As-Sahaba lil-Turath. find
  32. Taylor, P. and Black, A.W. 1998. „Assigning Phrase-Breaks from Part-of-Speech Sequences.‟ In Computer Speech and Language. 12.2: 99-117. ↗
  33. Taylor, L.J. and Knowles, G. 1988. ‘Manual of Information to Accompany the SEC Corpus: The machine readable corpus of spoken English.‟ Accessed: January 2010. http://khnt.hit.uib.no/icame/manuals/sec/INDEX.HTM ↗
  34. Wright, W. 1996. A Grammar of the Arabic Language, Translated from the German of Caspari, and Editted with Numerous Additions and Corrections Beirut: Librairie du Liban. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/brierley-2012-open-source-boundary-annotated-corpus/