--- title: "Al-Bayan — paper by Abdelnasser et al. (2014)" description: "To build this, the authors assembled several resources: a merged and manually… 0.847 precision. 12 references, cited by 16 papers here." authors: ["Heba Abdelnasser", "Maha Ragab", "Reham Mohamed", "Alaa Mohamed", "Bassant Farouk", "Nagwa El-Makky", "Marwan Torki"] year: 2014 venue: "Proceedings of the EMNLP 2014 Workshop on Arabic Natural Language Processing (ANLP)" source: "https://aclanthology.org/W14-3607/" doi: "10.3115/v1/w14-3607" licence: "CC BY-NC-SA 3.0 (ACL Anthology, before 2016)" licence_url: "https://creativecommons.org/licenses/by-nc-sa/3.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/abdelnasser-2014-al-bayan-arabic-question-answering/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/abdelnasser-2014-al-bayan-arabic-question-answering/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2014](https://md.research.elamur.ai/free/en/papers/2014/page/2/) / Al-Bayan: An Arabic Question Answering System for the Holy Quran # Al-Bayan: An Arabic Question Answering System for the Holy Quran **Authors:** [Heba Abdelnasser](https://md.research.elamur.ai/free/en/authors/heba-abdelnasser/), [Maha Ragab](https://md.research.elamur.ai/free/en/authors/maha-ragab/), [Reham Mohamed](https://md.research.elamur.ai/free/en/authors/reham-mohamed/), Alaa Mohamed, Bassant Farouk, [Nagwa El-Makky](https://md.research.elamur.ai/free/en/authors/nagwa-el-makky/) and [Marwan Torki](https://md.research.elamur.ai/free/en/authors/marwan-torki/) · **Year:** 2014 · **Venue:** [Proceedings of the EMNLP 2014 Workshop on Arabic Natural Language Processing (ANLP)](https://md.research.elamur.ai/free/en/venues/emnlp-workshop-on-arabic-natural-language-processing-anlp/) ([ANLP 2014](https://md.research.elamur.ai/free/en/venues/emnlp-workshop-on-arabic-natural-language-processing-anlp/2014/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY-NC-SA 3.0 (ACL Anthology, before 2016)](https://creativecommons.org/licenses/by-nc-sa/3.0/) **Links:** [Publisher's page](https://aclanthology.org/W14-3607/) · [PDF](https://aclanthology.org/W14-3607.pdf) · [DOI 10.3115/v1/w14-3607](https://doi.org/10.3115/v1/w14-3607) · [OpenAlex](https://openalex.org/W2251102620) **Citations:** 70 citations (OpenAlex, 2026-10-14) · [16](https://md.research.elamur.ai/free/en/papers/abdelnasser-2014-al-bayan-arabic-question-answering/cited-by/) in our library, of the 149,439 papers whose bibliographies we read ## Summary Al-Bayan is a question answering system for Arabic questions about the Holy Quran. The authors argue that Arabic QA is still weak and that religious topics need trustworthy answers, since the web holds much unreliable content about them. The system draws only on the Quran and two classical interpretation books (Tafseer). It works in three stages: - **Question analysis.** MADA performs morphological analysis and disambiguation, and an SVM assigns each question an expected answer type. - **Retrieval.** A variant of explicit semantic analysis maps text onto a hand-curated Quranic concept ontology instead of Wikipedia. - **Answer extraction.** The answer passage is taken from the retrieved verses and their Tafseer, ranked with features that include named entities. To build this, the authors assembled several resources: a merged and manually revised ontology covering all 6236 verses, a Quran-specific NER corpus, and a small labelled set of questions. Five Quran experts judged 59 questions. The full system found a correct answer within its top 3 for about 85% of them. The work matters as an early QA effort for Classical (Quranic) Arabic and as an example of building a narrow domain from curated, trusted sources. ## Key points - **Preprocessing.** MADA reduces every term to a (stem:POS) pair in Buckwalter transliteration. Words in function POS classes are dropped as stopwords, which matters because the Quran is fully diacritized while user questions usually are not. - **Question classification.** - The taxonomy has six answer types: Creation, Entity, Physical, Location, Number and Description. - The data is 230 questions, split 180 for training and 50 for testing. - The SVM scored 77.2% in 3-fold cross-validation and 86% on the 50 held-out questions. - **Ontology.** - It combines the Qurany concept tree with the Quranic Arabic Corpus ontology, and the 1,200 concepts were reviewed by hand. - 621 verses had no concept and were placed using a similarity module. - The final ontology has 1217 leaf concepts. - Retrieval uses Lucene with TFIDF weighting over the verses plus the Tafseer of Ibn-Kathir and Al-Jaza'iri. - **NER.** - Roughly 14 Quranic Arabic Corpus classes were collapsed into 5, and a Numbers class was tagged with help from a reference book. - The data follows the CoNLL 2002 BIO format. - Evaluation used LingPipe with 3 folds. - **Expert evaluation.** - IR module: Top-1 0.692, Top-5 0.847, precision 0.57. - Overall system: Top-1 0.650, Top-3 0.854, precision 0.73. ## Interesting ideas - The answer type is learned from the full question, not keyed on the wh-word, so a question asking for a mountain's name still comes out as Location. - When the main text lacks the answer, the system falls back on trusted commentary attached to each verse, as in the Queen of Sheba example where the name comes only from the Tafseer. - The answer-type taxonomy is deliberately aligned with the NER classes, plus one class for definitional questions, so classifier output maps directly onto candidate entities. - Items that no source had classified were attached to the concept of their most similar classified neighbours, completing ontology coverage. - Each module was evaluated separately, and an independent held-out set was reported alongside cross-validation. The end-to-end system was judged by several domain experts using TopN accuracy and precision. ## For Mova How these ideas could serve Mova, our small open language model — its files, stages and tests: - **I1** (answer): Our readers could derive the expected answer type from the head noun of the questioned phrase in the UD tree, using its ontology class, instead of relying only on who/where/when. - **I2** (world): Each entity in the world of a text could carry a linked layer of vetted background knowledge, consulted only after the text itself fails and always cited as an external source. - **I3** (answer): Defining question types in terms of our ontology's entity classes would let the answer step filter world entities by type in one explainable move. - **I4** (world): Verbs or nouns missing from VerbNet/WordNet-derived classes could get a provisional class from similar, already-classified items, flagged as inferred so the step stays explainable. - **I5** (learn): Our evaluation could report per-stage scores plus a fresh held-out set next to cross-validated numbers, which would show whether gains are real or come from overfitting a small set. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* 📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works · [🗄️ Information Systems](https://md.research.elamur.ai/free/en/shelves/computer-science/information-systems/) · 1,437 works 🧭 Topics (2): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 2,908 works · [❓ Question answering](https://md.research.elamur.ai/free/en/topics/question-answering/) · 441 works ## Cite Cite as: Heba Abdelnasser, Maha Ragab, Reham Mohamed, Alaa Mohamed, Bassant Farouk, Nagwa El-Makky and Marwan Torki. 2014. Al-Bayan: An Arabic Question Answering System for the Holy Quran. Proceedings of the EMNLP 2014 Workshop on Arabic Natural Language Processing (ANLP). The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{abdelnasser2014al, title = {{Al-Bayan: An Arabic Question Answering System for the Holy Quran}}, author = {Heba Abdelnasser, Maha Ragab, Reham Mohamed, Alaa Mohamed, Bassant Farouk, Nagwa El-Makky and Marwan Torki}, booktitle = {Proceedings of the EMNLP 2014 Workshop on Arabic Natural Language Processing (ANLP)}, year = {2014}, doi = {10.3115/v1/w14-3607}, url = {https://doi.org/10.3115/v1/w14-3607} } ``` - [← previous Epistemic modals and context: Experimental data](https://md.research.elamur.ai/free/en/papers/knobe-2014-epistemic-modals-context-experimental-data/) - [next → PropBank: Semantics of New Predicate Types](https://md.research.elamur.ai/free/en/papers/bonial-2014-propbank-semantics-new-predicate-types/) - [↑ back to papers of 2014](https://md.research.elamur.ai/free/en/papers/2014/page/2/) ## References 12 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Abdul-Baquee M. Sharaf and Eric Atwell. 2012. [QurAna: Corpus of the Quran annotated with Pronominal Anaphora](https://md.research.elamur.ai/free/en/papers/sharaf-2012-qurana-corpus-quran-annotated-pronominal/). LREC. 2. Abdul-Baquee M. Sharaf and Eric Atwell. 2012. Qur-Sim: A corpus for evaluation of relatedness in short texts. LREC. [find](https://scholar.google.com/scholar?q=Abdul%2DBaquee+M%2E+Sharaf+and+Eric+Atwell%2E+2012%2E+Qur%2DSim%3A+A+corpus+for+evaluation+of+relatedness+in+short+texts%2E+LREC%2E) 3. Abraham Ittycheriah and Salim Roukos. 2006. IBM’s statistical question answering system-TREC-11. Technical report, DTIC Document. [↗](https://doi.org/10.21236/ada456310) 4. Abu Bakr Al-Jaza’iri. 1986. Kalaam il ’Aliyy il Kabir. Aysar al-Tafasir li Abu Islam Ahmed bin Ali. 2008. Numbers and Ratios in the Quran. [find](https://scholar.google.com/scholar?q=Abu+Bakr+Al%2DJaza%E2%80%99iri%2E+1986%2E+Kalaam+il+%E2%80%99Aliyy+il+Kabir%2E+Aysar+al%2DTafasir+li+Abu+Islam+Ahmed+bin+Ali%2E+2008%2E+Numbers+and+Ratios+in+the+Quran%2E) 5. Bassam Hammo, Hani Abu-Salem and Steven Lytinen. 2002. [QARAB: a question answering system to support the Arabic language](https://md.research.elamur.ai/free/en/papers/hammo-2002-qarab-question-answering-system-support/). Proceedings of the ACL-02 workshop on Computational approaches to semitic languages. 6. Christopher D Manning, Prabhakar Raghavan and Hinrich Schütze. 2008. Introduction to information retrieval, volume 1. Cambridge university press Cambridge. [find](https://scholar.google.com/scholar?q=Christopher+D+Manning%2C+Prabhakar+Raghavan+and+Hinrich+Schu%CC%88tze%2E+2008%2E+Introduction+to+information+retrieval%2C+volume+1%2E+Cambridge+university+press+Cambridge%2E) 7. Eric Atwell, Claire Brierley, Kais Dukes, Majdi Sawalha and Abdul-Baquee Sharaf. 2011. A An Artificial Intelligence Approach to Arabic and Islamic Content on the Internet. Proceedings of NITS 3rd National Information Technology Symposium. [find](https://scholar.google.com/scholar?q=Eric+Atwell%2C+Claire+Brierley%2C+Kais+Dukes%2C+Majdi+Sawalha+and+Abdul%2DBaquee+Sharaf%2E+2011%2E+A+An+Artificial+Intelligence+Approach+to+Arabic+and+Islamic+Content+on+the+Internet%2E+Proceedings+of+NITS+3rd+Nati) 8. Evgeniy Gabrilovich and Shaul Markovitch. 2007. Computing Semantic Relatedness Using Wikipedia-based Explicit Semantic Analysis, volume 7. Proceedings of the 20th international joint conference on artificial intelligence. [find](https://scholar.google.com/scholar?q=Evgeniy+Gabrilovich+and+Shaul+Markovitch%2E+2007%2E+Computing+Semantic+Relatedness+Using+Wikipedia%2Dbased+Explicit+Semantic+Analysis%2C+volume+7%2E+Proceedings+of+the+20th+international+joint+conference+on+art) 9. Hikmat Ullah Khan and Syed Muhammad Saqlain and Muhammad Shoaib and Muhammad Sher. 2013. Ontology Based Semantic Search in Holy Quran., volume 2. International Journal of Future Computer and Communication, 570-575. Ismail Ibn-Kathir. 1370. Tafsir al-Qur’an al-Azim. [find](https://scholar.google.com/scholar?q=Hikmat+Ullah+Khan+and+Syed+Muhammad+Saqlain+and+Muhammad+Shoaib+and+Muhammad+Sher%2E+2013%2E+Ontology+Based+Semantic+Search+in+Holy+Quran%2E%2C+volume+2%2E+International+Journal+of+Future+Computer+and+Communica) 10. Manal AlMaayah, Majdi Sawalha, and Mohammad AM Abushariah. 2014. A Proposed Model for Quranic Arabic WordNet. LRE-REL2, 9. [find](https://scholar.google.com/scholar?q=Manal+AlMaayah%2C+Majdi+Sawalha%2C+and+Mohammad+AM+Abushariah%2E+2014%2E+A+Proposed+Model+for+Quranic+Arabic+WordNet%2E+LRE%2DREL2%2C+9%2E) 11. Mengqiu Wang. 2006. A Survey of Answer Extraction Techniques in Factoid Question Answering, volume 1. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Mengqiu+Wang%2E+2006%2E+A+Survey+of+Answer+Extraction+Techniques+in+Factoid+Question+Answering%2C+volume+1%2E+Association+for+Computational+Linguistics%2E) 12. Michael Fleischman, Eduard Hovy and Abdessamad Echihabi. 2003. Offline strategies for online question answering: answering questions before they are asked, volume 1. Proceedings of the 41st Annual Meeting on Association for Computational Linguistics. ؟ أبس ةكلم يه نم [find](https://scholar.google.com/scholar?q=Michael+Fleischman%2C+Eduard+Hovy+and+Abdessamad+Echihabi%2E+2003%2E+Offline+strategies+for+online+question+answering%3A+answering+questions+before+they+are+asked%2C+volume+1%2E+Proceedings+of+the+41st+Annual+Mee) *Updated 1 November 2026*