---
title: "Cross-Dialectal Named Entity Recognition in Arabic — Elkhbir 2023"
description: "Most annotated Arabic NER data covers Modern Standard Arabic (MSA). 21 references, cited by 3 papers here."
authors: ["Niama Elkhbir", "Urchade Zaratiana", "Nadi Tomeh", "Thierry Charnois"]
year: 2023
venue: "Proceedings of ArabicNLP 2023"
source: "https://aclanthology.org/2023.arabicnlp-1.12/"
doi: "10.18653/v1/2023.arabicnlp-1.12"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/elkhbir-2023-cross-dialectal-named-entity-recognition/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/elkhbir-2023-cross-dialectal-named-entity-recognition/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2023](https://md.research.elamur.ai/free/en/papers/2023/page/12/) / Cross-Dialectal Named Entity Recognition in Arabic

# Cross-Dialectal Named Entity Recognition in Arabic

**Authors:** [Niama Elkhbir](https://md.research.elamur.ai/free/en/authors/niama-elkhbir/), [Urchade Zaratiana](https://md.research.elamur.ai/free/en/authors/urchade-zaratiana/), [Nadi Tomeh](https://md.research.elamur.ai/free/en/authors/nadi-tomeh/) and [Thierry Charnois](https://md.research.elamur.ai/free/en/authors/thierry-charnois/) · **Year:** 2023 · **Venue:** [Proceedings of ArabicNLP 2023](https://md.research.elamur.ai/free/en/venues/arabicnlp/) ([ArabicNLP 2023](https://md.research.elamur.ai/free/en/venues/arabicnlp/2023/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2023.arabicnlp-1.12/) · [PDF](https://aclanthology.org/2023.arabicnlp-1.12.pdf) · [DOI 10.18653/v1/2023.arabicnlp-1.12](https://doi.org/10.18653/v1/2023.arabicnlp-1.12) · [OpenAlex](https://openalex.org/W4389518283)

**Citations:** 4 citations (OpenAlex, 2026-10-14) · [3](https://md.research.elamur.ai/free/en/papers/elkhbir-2023-cross-dialectal-named-entity-recognition/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
Most annotated Arabic NER data covers Modern Standard Arabic (MSA). Dialects such as Moroccan, Egyptian and Syrian have little or no labeled data, no fixed spelling, and different morphology and syntax. The authors ask how well an MSA-trained recognizer works on dialect text with no dialect annotation. They took 500 MSA sentences from ACE 2005 and kept only names and nominal mentions as flat entities. They also hand-labeled about 500 sentences per dialect from the xP3x corpus with the seven ACE types. A span-based classifier was fine-tuned on several Arabic and multilingual encoders and tested zero-shot across varieties, in both directions. Zero-shot transfer gave usable accuracy at no annotation cost. Syrian was closest to MSA, Egyptian was also strong, and Moroccan was consistently the hardest. The annotations and models are public, so they provide a starting benchmark for dialectal NER.

## Key points
- MSA side: 500 random sentences from ACE 2005 (14,168 tokens, 3,030 entities), plus 350 extra MSA sentences used to train an in-variety reference model.
- Dialect side: sentences from xP3x were split on whitespace and labeled in Label Studio by one annotator, a Moroccan speaker who also knows Egyptian and Syrian. Only sentences with at least one entity were kept:
  - Moroccan: 378 sentences, 970 entities.
  - Egyptian: 353 sentences, 831 entities.
  - Syrian: 361 sentences, 956 entities.
- The annotation scheme comes from the ACE/LDC guidelines and uses seven categories: person (PER), organization (ORG), geopolitical entity (GPE), location (LOC), facility (FAC), vehicle (VEH) and weapon (WEA). Only NAM and NOM mentions are included; pronouns (PRO), nesting and coreference are excluded, and the innermost flat entity is annotated.
- NER is cast as classifying every candidate span. Training sums a per-span loss, so the global partition function is never computed.
- Encoders compared: MARBERTv2, AraBERTv2, CAMeLBERT-DA, CAMeLBERT-Mix, CAMeLBERT-MSA, mBERT and mDeBERTav3. In the larger-MSA-data experiment, AraBERTv2 and CAMeLBERT-MSA scored highest and CAMeLBERT-DA lowest.

## Interesting ideas
- Running transfer in both directions (source→target and target→source) gives a cheap measure of how close two varieties are, not just a single accuracy score.
- Annotating only the smallest flat constituent keeps the guidelines simple and the labels portable across varieties; nesting is postponed rather than guessed.
- Pronoun mentions were dropped because they depend heavily on context and are ambiguous across varieties, leaving them for later work instead of adding label noise.
- Taking evaluation sentences from a parallel translation corpus makes the dialect test sets similar in content, so score gaps reflect language variety more than topic.
- A small in-variety model trained on extra data gives a reference ceiling against which zero-shot transfer is judged.

## For Mova

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

- **I1** (tree): train gparse on one English treebank or genre, test on the others and in reverse, to map which UD sources are close enough to merge.
- **I2** (world): entity extraction could first emit minimal flat mentions and build larger composite entities in a separate, explainable layer.
- **I3** (frames): keep pronoun handling strictly in the coreference stage, so mention detection is scored only on names and nominals.
- **I4** (learn): build test sets from parallel or paraphrased sentences, so variance from register or wording is separated from content variance.
- **I5** (learn): pair every silver-data or transfer evaluation with an in-domain reference run, so gains can be read against a known upper bound.

*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

- [Centre National de la Recherche Scientifique](https://md.research.elamur.ai/free/en/labs/centre-national-de-la-recherche-scientifique/) — Research institute, Paris, France · 168 works · 62,369 citations
- [Laboratoire d'Informatique de Paris-Nord](https://md.research.elamur.ai/free/en/labs/laboratoire-d-informatique-de-paris-nord/) — Research institute, Villetaneuse, France · 9 works · 510 citations

📚 Shelves (4): [🖥️ 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

🧭 Topics (2): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 2,908 works · [🏷️ Named entity recognition](https://md.research.elamur.ai/free/en/topics/named-entity-recognition/) · 200 works

## Cite

Cite as: Niama Elkhbir, Urchade Zaratiana, Nadi Tomeh and Thierry Charnois. 2023. Cross-Dialectal Named Entity Recognition in Arabic. Proceedings of ArabicNLP 2023.

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

```bibtex
@inproceedings{elkhbir2023cross,
  title = {{Cross-Dialectal Named Entity Recognition in Arabic}},
  author = {Niama Elkhbir, Urchade Zaratiana, Nadi Tomeh and Thierry Charnois},
  booktitle = {Proceedings of ArabicNLP 2023},
  year = {2023},
  doi = {10.18653/v1/2023.arabicnlp-1.12},
  url = {https://doi.org/10.18653/v1/2023.arabicnlp-1.12}
}
```

- [← previous Are UD Treebanks Getting More Consistent? A Report Card for English UD](https://md.research.elamur.ai/free/en/papers/zeldes-2023-are-ud-treebanks-getting-more/)
- [next → Explicit Morphological Knowledge Improves Pre-training of Language Models for Hebrew](https://md.research.elamur.ai/free/en/papers/gueta-2023-explicit-morphological-knowledge-improves-pre/)
- [↑ back to papers of 2023](https://md.research.elamur.ai/free/en/papers/2023/page/12/)

## References

21 references, 13 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Pengcheng He, Xiaodong Liu, Jianfeng Gao, and Weizhu Chen. 2020. [Deberta: Decoding-enhanced bert with disentangled attention](https://md.research.elamur.ai/free/en/papers/he-2020-deberta-decoding-enhanced-bert-disentangled/). ArXiv, abs/2006.03654. [↗](https://arxiv.org/abs/2006.03654)
2. Sherief Abdallah, Khaled Shaalan, and Muhammad Shoaib. 2012. Integrating rule-based system with classification for arabic named entity recognition. In Computational Linguistics and Intelligent Text Processing, pages 311–322, Berlin, Heidelberg. Springer Berlin Heidelberg. [find](https://scholar.google.com/scholar?q=Sherief+Abdallah%2C+Khaled+Shaalan%2C+and+Muhammad+Shoaib%2E+2012%2E+Integrating+rule%2Dbased+system+with+classification+for+arabic+named+entity+recognition%2E+In+Computational+Linguistics+and+Intelligent+Text+Pr)
3. Chadi Helwe, Ghassan Dib, Mohsen Shamas, and Shady Elbassuoni. 2020. [A semi-supervised BERT approach for Arabic named entity recognition](https://md.research.elamur.ai/free/en/papers/helwe-2020-semi-supervised-bert-approach-arabic/). In Proceedings of the Fifth Arabic Natural Language Processing Workshop, pages 49–57, Barcelona, Spain (Online). Association for Computational Linguistics.
4. Muhammad Abdul-Mageed, AbdelRahim Elmadany, and El Moatez Billah Nagoudi. 2021. [ARBERT & MARBERT: Deep bidirectional transformers for Arabic](https://md.research.elamur.ai/free/en/papers/abdul-mageed-2021-arbert-marbert-deep-bidirectional-transformers/). In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 7088–7105, Online. Association for Computational Linguistics. [↗](https://doi.org/10.18653/v1/2021.acl-long.551)
5. Go Inoue, Bashar Alhafni, Nurpeiis Baimukan, Houda Bouamor, and Nizar Habash. 2021. [The interplay of variant, size, and task type in Arabic pre-trained language models](https://md.research.elamur.ai/free/en/papers/inoue-2021-interplay-variant-size-task-type/). In Proceedings of the Sixth Arabic Natural Language Processing Workshop, Kyiv, Ukraine (Online). Association for Computational Linguistics.
6. Muhammad Saleh Al-Qurishi and Riad Souissi. 2021. [Arabic named entity recognition using transformer-based-CRF model](https://md.research.elamur.ai/free/en/papers/qurishi-2021-arabic-named-entity-recognition-transformer/). In Proceedings of the 4th International Conference on Natural Language and Speech Processing (ICNLSP 2021), pages 262–271, Trento, Italy. Association for Computational Linguistics.
7. Wissam Antoun, Fady Baly, and Hazem Hajj. 2020. [AraBERT: Transformer-based model for Arabic language understanding](https://md.research.elamur.ai/free/en/papers/antoun-2020-arabert-transformer-based-model-arabic/). In Proceedings of the 4th Workshop on Open-Source Arabic Corpora and Processing Tools, with a Shared Task on Offensive Language Detection, pages 9–15, Marseille, France. European Language Resource Association. [↗](https://doi.org/10.48550/arxiv.2003.00104)
8. Wissam Antoun, Fady Baly, and Hazem Hajj. 2021. [Araelectra: Pre-training text discriminators for arabic language understanding](https://md.research.elamur.ai/free/en/papers/antoun-2021-araelectra-pre-training-text-discriminators/). [↗](https://doi.org/10.48550/arxiv.2012.15516)
9. Mustafa Jarrar, Mohammed Khalilia, and Sana Ghanem. 2022. [Wojood: Nested Arabic named entity corpus and recognition using BERT](https://md.research.elamur.ai/free/en/papers/jarrar-2022-wojood-nested-arabic-named-entity/). In Proceedings of the Thirteenth Language Resources and Evaluation Conference, pages 3626–3636, Marseille, France. European Language Resources Association.
10. Mohamed A. Meselhi, Hitham M. Abo Bakr, Ibrahim Ziedan, and Khaled Shaalan. 2014. A novel hybrid approach to arabic named entity recognition. In Machine Translation, pages 93–103, Berlin, Heidelberg. Springer Berlin Heidelberg. [find](https://scholar.google.com/scholar?q=Mohamed+A%2E+Meselhi%2C+Hitham+M%2E+Abo+Bakr%2C+Ibrahim+Ziedan%2C+and+Khaled+Shaalan%2E+2014%2E+A+novel+hybrid+approach+to+arabic+named+entity+recognition%2E+In+Machine+Translation%2C+pages+93%E2%80%93103%2C+Berlin%2C+Heidelberg%2E+)
11. Behrang Mohit, Nathan Schneider, Rishav Bhowmick, Kemal Oflazer, and Noah A. Smith. 2012. [Recall-oriented learning of named entities in Arabic Wikipedia](https://md.research.elamur.ai/free/en/papers/mohit-2012-recall-oriented-learning-named-entities/). In Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics, pages 162–173, Avignon, France. Association for Computational Linguistics.
12. Hanane Nour Moussa and Asmaa Mourhir. 2023. Darnercorp: An annotated named entity recognition dataset in the moroccan dialect. Data in Brief, 48:109234. [↗](https://doi.org/10.60692/dhywt-y6a84)
13. Niklas Muennighoff, Thomas Wang, Lintang Sutawika, Adam Roberts, Stella Biderman, Teven Le Scao, M Saiful Bari, Sheng Shen, Zheng-Xin Yong, Hailey Schoelkopf, et al. 2022. [Crosslingual generalization through multitask finetuning](https://md.research.elamur.ai/free/en/papers/muennighoff-2022-crosslingual-generalization-through-multitask-finetuning/). arXiv preprint arXiv:2211.01786. [↗](https://arxiv.org/abs/2211.01786)
14. Mai Oudah and Khaled Shaalan. 2012. [A pipeline Arabic named entity recognition using a hybrid approach](https://md.research.elamur.ai/free/en/papers/oudah-2012-pipeline-arabic-named-entity-recognition/). In Proceedings of COLING 2012, pages 2159–2176, Mumbai, India.
15. The COLING 2012 Organizing Committee. [find](https://scholar.google.com/scholar?q=The+COLING+2012+Organizing+Committee%2E)
16. Khaled Shaalan and Hafsa Raza. 2008. Arabic named entity recognition from diverse text types. In Advances in Natural Language Processing, pages 440– 451, Berlin, Heidelberg. Springer Berlin Heidelberg. [find](https://scholar.google.com/scholar?q=Khaled+Shaalan+and+Hafsa+Raza%2E+2008%2E+Arabic+named+entity+recognition+from+diverse+text+types%2E+In+Advances+in+Natural+Language+Processing%2C+pages+440%E2%80%93+451%2C+Berlin%2C+Heidelberg%2E+Springer+Berlin+Heidelberg)
17. C. Walker and Linguistic Data Consortium. 2005. ACE 2005 Multilingual Training Corpus. LDC corpora. Linguistic Data Consortium. [find](https://scholar.google.com/scholar?q=C%2E+Walker+and+Linguistic+Data+Consortium%2E+2005%2E+ACE+2005+Multilingual+Training+Corpus%2E+LDC+corpora%2E+Linguistic+Data+Consortium%2E)
18. Urchade Zaratiana, Niama Elkhbir, Pierre Holat, Nadi Tomeh, and Thierry Charnois. 2022a. Global span selection for named entity recognition. In Proceedings of the Workshop on Unimodal and Multimodal Induction of Linguistic Structures (UM-IoS), pages 11–17, Abu Dhabi, United Arab Emirates (Hybrid). Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Urchade+Zaratiana%2C+Niama+Elkhbir%2C+Pierre+Holat%2C+Nadi+Tomeh%2C+and+Thierry+Charnois%2E+2022a%2E+Global+span+selection+for+named+entity+recognition%2E+In+Proceedings+of+the+Workshop+on+Unimodal+and+Multimodal+I)
19. Urchade Zaratiana, Nadi Tomeh, Pierre Holat, and Thierry Charnois. 2022b. Named entity recognition as structured span prediction. In Proceedings of the Workshop on Unimodal and Multimodal Induction of Linguistic Structures (UM-IoS), pages 1–10, Abu Dhabi, United Arab Emirates (Hybrid). Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Urchade+Zaratiana%2C+Nadi+Tomeh%2C+Pierre+Holat%2C+and+Thierry+Charnois%2E+2022b%2E+Named+entity+recognition+as+structured+span+prediction%2E+In+Proceedings+of+the+Workshop+on+Unimodal+and+Multimodal+Induction+of)
20. Ayah Zirikly and Mona Diab. 2014. [Named entity recognition system for dialectal Arabic](https://md.research.elamur.ai/free/en/papers/zirikly-2014-named-entity-recognition-system-dialectal/). In Proceedings of the EMNLP 2014 Workshop on Arabic Natural Language Processing (ANLP), pages 78–86, Doha, Qatar. Association for Computational Linguistics.
21. Ayah Zirikly and Mona Diab. 2015. [Named entity recognition for Arabic social media](https://md.research.elamur.ai/free/en/papers/zirikly-2015-named-entity-recognition-arabic-social/). In Proceedings of the 1st Workshop on Vector Space Modeling for Natural Language Processing, pages 176–185, Denver, Colorado. Association for Computational Linguistics.

*Updated 1 November 2026*