---
title: "CV-18 NER — paper by Saidi et al. (2026) — summary and ideas"
description: "Arabic has had no open benchmark for recognizing named entities directly from speech. 22 references."
authors: ["Youssef Saidi", "Haroun Elleuch", "Fethi Bougares"]
year: 2026
venue: "The 7th Workshop on Open-Source Arabic Corpora and Processing Tools (OSACT7) with 5 Shared Tasks"
source: "https://aclanthology.org/2026.osact-1.15/"
doi: "10.63317/3ayuttvcg6vu"
licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)"
licence_url: "https://creativecommons.org/licenses/by-nc/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/saidi-2026-cv-18-ner-augmented-common/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/saidi-2026-cv-18-ner-augmented-common/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/26/) / CV-18 NER: Augmented Common Voice for Named Entity Recognition from Arabic Speech

# CV-18 NER: Augmented Common Voice for Named Entity Recognition from Arabic Speech

**Authors:** [Youssef Saidi](https://md.research.elamur.ai/free/en/authors/youssef-saidi/), [Haroun Elleuch](https://md.research.elamur.ai/free/en/authors/haroun-elleuch/) and [Fethi Bougares](https://md.research.elamur.ai/free/en/authors/fethi-bougares/) · **Year:** 2026 · **Venue:** [The 7th Workshop on Open-Source Arabic Corpora and Processing Tools (OSACT7) with 5 Shared Tasks](https://md.research.elamur.ai/free/en/venues/osact/) ([Workshop on Open-Source Arabic Corpora and Processing Tools with Shared Tasks 2026](https://md.research.elamur.ai/free/en/venues/osact/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.osact-1.15/) · [PDF](https://aclanthology.org/2026.osact-1.15.pdf) · [DOI 10.63317/3ayuttvcg6vu](https://doi.org/10.63317/3ayuttvcg6vu) · [OpenAlex](https://openalex.org/W7166581155)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
Arabic has had no open benchmark for recognizing named entities directly from speech. In English, French and Chinese, end-to-end (E2E) systems have already been shown to beat cascaded ASR + text-NER pipelines, but Arabic has lagged behind.

The authors built CV-18 NER from the Modern Standard Arabic part of Common Voice 18, which is CC0-licensed. They added entity labels in BIO format using the 21-type Wojood tagset:
- **Pre-annotation:** several Arabic text encoders were fine-tuned on Wojood, and the best one, AraBERT v2, pre-labelled the transcripts.
- **Correction:** a single annotator then corrected these labels.
- **Filtering:** utterances without any entity were dropped.

They then compared cascaded systems with E2E models based on Whisper and on the Arabic self-supervised model AraBEST-RQ. The Whisper model is trained on transcripts with inline entity tags and an adapted tokenizer.

E2E models clearly outperformed the best pipeline on the test set. Arabic-only SSL pretraining helped transcription most. Whisper's multilingual weak supervision transferred better to joint speech-to-entity learning. Larger models were not necessarily easier to adapt with this little data.

The dataset and models are released on Hugging Face (Elyadata/CV18-NER).

## Key points
- **Dataset size:** filtering removed about 75% of utterances. The final splits are:

  | Split | Utterances | Duration (h:mm) |
  |---|---|---|
  | Train | 7,119 | 8:15 |
  | Dev | 2,263 | 2:54 |
  | Test | 2,325 | 2:59 |

- **Best test results:** 37.0% CoER with AraBEST-RQ 300M and 38.0% CVER with Whisper-medium.
- **Effect on transcription:** entity-aware training lowered Whisper-medium's WER by 1.3 absolute points compared with its plain ASR version.
- **Models compared:**
  - Whisper-medium and Whisper-large-v3.
  - AraBEST-RQ 300M (about 6k h of Arabic speech) and 600M (6k h, or 14k h including Common Voice 16).
  - Text NER models AraBERT v2, AraBERTv0.2, CAMeLbert-MSA and CAMeLbert-MIX.
- **Class imbalance and limitations:**
  - PERS dominates, with 5,314 tags in the training set.
  - LAW, CURR, PERCENT and PRODUCT are rare, and UNIT does not occur at all.
  - Because only one annotator was used, no inter-annotator agreement is reported.

*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 (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: Youssef Saidi, Haroun Elleuch and Fethi Bougares. 2026. CV-18 NER: Augmented Common Voice for Named Entity Recognition from Arabic Speech. The 7th Workshop on Open-Source Arabic Corpora and Processing Tools (OSACT7) with 5 Shared Tasks.

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

```bibtex
@inproceedings{saidi2026cv,
  title = {{CV-18 NER: Augmented Common Voice for Named Entity Recognition from Arabic Speech}},
  author = {Youssef Saidi, Haroun Elleuch and Fethi Bougares},
  booktitle = {The 7th Workshop on Open-Source Arabic Corpora and Processing Tools (OSACT7) with 5 Shared Tasks},
  year = {2026},
  doi = {10.63317/3ayuttvcg6vu},
  url = {https://doi.org/10.63317/3ayuttvcg6vu}
}
```

- [← previous Curvature aware gradient projection prototype meta learning framework improves low resource English dialect recognition](https://md.research.elamur.ai/free/en/papers/chen-2026-curvature-aware-gradient-projection-prototype/)
- [next → CVRH: Cross-modal Variational Role Hypergraph Network via Semantic Enhancement for Multi-modal Event Argument Extraction](https://md.research.elamur.ai/free/en/papers/pan-2026-cvrh-cross-modal-variational-role/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/26/)

## References

22 references, 9 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. 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 Resources Association. [↗](https://doi.org/10.48550/arxiv.2003.00104)
2. Rosana Ardila, Megan Branson, Kelly Davis, Michael Kohler, Josh Meyer, Michael Henretty, Reuben Morais, Lindsay Saunders, Francis Tyers, and Gregor Weber. 2020. [Common voice: A massively-multilingual speech corpus](https://md.research.elamur.ai/free/en/papers/ardila-2019-common-voice-massively-multilingual-speech/). In LREC, pages 4218–4222, Marseille, France. European Language Resources Association. [↗](https://doi.org/10.48550/arxiv.1912.06670)
3. Gil Ayache, Menachem Pirchi, Aviv Navon, Aviv Shamsian, Gill Hetz, and Joseph Keshet. 2025. Whisperner: Unified open named entity and speech recognition. [find](https://scholar.google.com/scholar?q=Gil+Ayache%2C+Menachem+Pirchi%2C+Aviv+Navon%2C+Aviv+Shamsian%2C+Gill+Hetz%2C+and+Joseph+Keshet%2E+2025%2E+Whisperner%3A+Unified+open+named+entity+and+speech+recognition%2E)
4. Yassine Benajiba, Paolo Rosso, and José-Miguel Benedí. 2009. Anersys: An arabic named entity recognition system based on maximum entropy. In Conference on Intelligent Text Processing and Computational Linguistics. [find](https://scholar.google.com/scholar?q=Yassine+Benajiba%2C+Paolo+Rosso%2C+and+Jos%C3%A9%2DMiguel+Bened%C3%AD%2E+2009%2E+Anersys%3A+An+arabic+named+entity+recognition+system+based+on+maximum+entropy%2E+In+Conference+on+Intelligent+Text+Processing+and+Computational)
5. Antoine Caubrière, Sophie Rosset, Yannick Estève, Antoine Laurent, and Emmanuel Morin. 2020. Where are we in named entity recognition from speech? In Proceedings of the Twelfth Language Resources and Evaluation Conference, pages 4514–4520, Marseille, France. European Language Resources Association. [find](https://scholar.google.com/scholar?q=Antoine+Caubri%C3%A8re%2C+Sophie+Rosset%2C+Yannick+Est%C3%A8ve%2C+Antoine+Laurent%2C+and+Emmanuel+Morin%2E+2020%2E+Where+are+we+in+named+entity+recognition+from+speech%3F+In+Proceedings+of+the+Twelfth+Language+Resources+and+)
6. Boli Chen, Guangwei Xu, Xiaobin Wang, Pengjun Xie, Meishan Zhang, and Fei Huang. 2022. Aishell-ner: Named entity recognition from chinese speech. In ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 8352–8356. [↗](https://doi.org/10.48550/arxiv.2202.08533)
7. 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](https://md.research.elamur.ai/free/en/papers/chowdhury-2021-towards-one-model-rule-all/). In Interspeech 2021, pages 2466–2470.
8. Charika de Alvis and Suranga Seneviratne. 2024. A survey of deep long-tail classification advancements. [↗](https://doi.org/10.48550/arxiv.2404.15593)
9. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. [BERT: Pre-training of deep bidirectional transformers for language understanding](https://md.research.elamur.ai/free/en/papers/devlin-2019-bert-pre-training-deep-bidirectional/). In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 4171–4186, Minneapolis, Minnesota. Association for Computational Linguistics. [↗](https://doi.org/10.18653/v1/n19-1423)
10. Haroun Elleuch, Youssef Saidi, Salima Mdhaffar, Yannick Estève, and Fethi Bougares. 2025. [Elyadata & lia at nadi 2025: Asr and adi subtasks](https://md.research.elamur.ai/free/en/papers/elleuch-2025-elyadata-lia-nadi-2025-asr/).
11. Haroun Elleuch, Ryan Whetten, Salima Mdhaffar, Yannick Estève, and Fethi Bougares. 2026. Arabest-rq: Multidialectal arabic ssl. In Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Accepted for publication. [find](https://scholar.google.com/scholar?q=Haroun+Elleuch%2C+Ryan+Whetten%2C+Salima+Mdhaffar%2C+Yannick+Est%C3%A8ve%2C+and+Fethi+Bougares%2E+2026%2E+Arabest%2Drq%3A+Multidialectal+arabic+ssl%2E+In+Proceedings+of+the+IEEE+International+Conference+on+Acoustics%2C+Speech)
12. 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 LREC, pages 3626–3636, Marseille, France. ELRA.
13. Yuang Li, Jiawei Yu, Min Zhang, Mengxin Ren, Yanqing Zhao, Xiaofeng Zhao, Shimin Tao, Jinsong Su, and Hao Yang. 2024. Using Large Language Model for End-to-End Chinese ASR and NER. In Interspeech 2024, pages 822–826. [find](https://scholar.google.com/scholar?q=Yuang+Li%2C+Jiawei+Yu%2C+Min+Zhang%2C+Mengxin+Ren%2C+Yanqing+Zhao%2C+Xiaofeng+Zhao%2C+Shimin+Tao%2C+Jinsong+Su%2C+and+Hao+Yang%2E+2024%2E+Using+Large+Language+Model+for+End%2Dto%2DEnd+Chinese+ASR+and+NER%2E+In+Interspeech+2024)
14. Salima Mdhaffar, Jarod Duret, Titouan Parcollet, and Yannick Estève. 2022. End-to-end model for named entity recognition from speech without paired training data. In Interspeech 2022, pages 4068–4072. [find](https://scholar.google.com/scholar?q=Salima+Mdhaffar%2C+Jarod+Duret%2C+Titouan+Parcollet%2C+and+Yannick+Est%C3%A8ve%2E+2022%2E+End%2Dto%2Dend+model+for+named+entity+recognition+from+speech+without+paired+training+data%2E+In+Interspeech+2022%2C+pages+4068%E2%80%934072%2E)
15. 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.
16. Sota Nemoto, Shunsuke Kitada, and Hitoshi Iyatomi. 2025. Majority or minority: Data imbalance learning method for named entity recognition. IEEE Access, 13:9902–9909. [↗](https://doi.org/10.1109/access.2024.3522972)
17. Alec Radford, Jong Wook Kim, Tao Xu, Greg Brockman, Christine Mcleavey, and Ilya Sutskever. 2023. [Robust speech recognition via large-scale weak supervision](https://md.research.elamur.ai/free/en/papers/radford-2022-robust-speech-recognition-large-scale/). In ICML, volume 202 of Proceedings of Machine Learning Research, pages 28492–28518. PMLR. [↗](https://doi.org/10.48550/arxiv.2212.04356)
18. Mirco Ravanelli, Titouan Parcollet, Adel Moumen, Sylvain De Langen, Cem Subakan, Peter Plantinga, Yingzhi Wang, Pooneh Mousavi, Luca Della Libera, Artem Ploujnikov, Francesco Paissan, Davide Borra, Salah Zaiem, Zeyu Zhao, Shucong Zhang, Georgios Karakasidis, Sung-Lin Yeh, Pierre Champion, Aku Rouhe, Rudolf Braun, Florian Mai, Juan Zuluaga-Gomez, Seyed Mahed Mousavi, Andreas Nautsch, Ha Nguyen, Xuechen Liu, Sangeet Sagar, Jarod Duret, Salima Mdhaffar, Gaëlle Laperrière, Mickael Rouvier, Renato De Mori, and Yannick Estève. 2024. Open-source conversational ai with speechbrain 1.0. J. Mach. Learn. Res., 25(1). [↗](https://doi.org/10.48550/arxiv.2407.00463)
19. Piotr Szymański, Lukasz Augustyniak, Mikolaj Morzy, Adrian Szymczak, Krzysztof Surdyk, and Piotr Żelasko. 2023. Why aren’t we NER yet? artifacts of ASR errors in named entity recognition in spontaneous speech transcripts. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1746–1761, Toronto, Canada. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Piotr+Szyma%C5%84ski%2C+Lukasz+Augustyniak%2C+Mikolaj+Morzy%2C+Adrian+Szymczak%2C+Krzysztof+Surdyk%2C+and+Piotr+%C5%BBelasko%2E+2023%2E+Why+aren%E2%80%99t+we+NER+yet%3F+artifacts+of+ASR+errors+in+named+entity+recognition+in+spontaneou)
20. Bashar Talafha, Karima Kadaoui, Samar Mohamed Magdy, Mariem Habiboullah, Chafei Mohamed Chafei, Ahmed Oumar El-Shangiti, Hiba Zayed, Mohamedou cheikh tourad, Rahaf Alhamouri, Rwaa Assi, Aisha Alraeesi, Hour Mohamed, Fakhraddin Alwajih, Abdelrahman Mohamed, Abdellah El Mekki, El Moatez Billah Nagoudi, Benelhadj Djelloul Mama Saadia, Hamzah A. Alsayadi, Walid Al-Dhabyani, Sara Shatnawi, Yasir Ech-Chammakhy, Amal Makouar, Yousra Berrachedi, Mustafa Jarrar, Shady Shehata, Ismail Berrada, and Muhammad Abdul-Mageed. 2024. [Casablanca: Data and models for multidialectal arabic speech recognition](https://md.research.elamur.ai/free/en/papers/talafha-2024-casablanca-data-models-multidialectal-arabic/).
21. Hemant Yadav, Sreyan Ghosh, Yi Yu, and Rajiv Ratn Shah. 2020. End-to-end named entity recognition from english speech. [find](https://scholar.google.com/scholar?q=Hemant+Yadav%2C+Sreyan+Ghosh%2C+Yi+Yu%2C+and+Rajiv+Ratn+Shah%2E+2020%2E+End%2Dto%2Dend+named+entity+recognition+from+english+speech%2E)
22. Rodolfo Zevallos, Mireia Farrús, and Núria Bel. 2023. Frequency balanced datasets lead to better language models. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 7859–7872, Singapore. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Rodolfo+Zevallos%2C+Mireia+Farr%C3%BAs%2C+and+N%C3%BAria+Bel%2E+2023%2E+Frequency+balanced+datasets+lead+to+better+language+models%2E+In+Findings+of+the+Association+for+Computational+Linguistics%3A+EMNLP+2023%2C+pages+7859%E2%80%937)

*Updated 1 November 2026*