AI Elites at NADI 2026 Shared Task: Selecting on Out-of-Domain Accuracy for Cross-Domain Spoken Arabic Dialect Identification

Abstract

We describe our submission to NADI-2026 Subtask 2, country-level spoken Arabic dialect identification on a blind test set from undisclosed sources. Because the target domain is unknown at development time, we select on out-of-domain accuracy rather than in-domain validation accuracy at every decision point, comparing three backbones under a common pooling-plus-linear head and training the selected Arabic ASR encoder on ADI20-micro mixed with Casablanca. The system reaches 91.54% in-domain, 73.10% on the held-out MADIS-5 corpus (scored five-way), and 52.51% on the blind test set (C _avg0.11), placing fourth. The criterion helped, choosing an encoder that in-domain accuracy would have rejected by 32 points on a held-out corpus, a gap more than twenty times our measured seed variation, but it was not sufficient: of the 39.03-point in-domain to blind gap, at least 18.44 points are attributable to leaving the training domain. We trace part of the remainder to a corpus that entered the training mixture while still being used for checkpoint selection.

- Anthology ID:
- 2026.arabicnlp-sharedtasks.10
- Volume:
- [Proceedings of The Fourth Arabic Natural Language Processing Conference: Shared Tasks](https://aclanthology.org/volumes/2026.arabicnlp-sharedtasks/)
- Month:

• October
• Year:
• 2026
• Address:
• Budapest, Hungary
• Editors:
• Muhammad Abdul-Mageed ,Tamer Elsayed ,Houda Bouamor ,Samhaa El-Beltagy ,Hend Al-Khalifa ,Ahmed Abdelali ,AbdelRahim Elmadany ,Abdellah El Mekki ,Kareem Darwish ,Sakhar Alkhereyf ,Hamzah Luqman ,Salam Khalifa ,Samar Magdy ,Zaid Alyafeai ,Injy Hamed ,Wajdi Zaghouani ,Imed Zitouni

- Venue:
- [ArabicNLP](https://aclanthology.org/venues/arabicnlp/)
- SIG:
- [SIGARAB](https://aclanthology.org/sigs/sigarab/)
- Publisher:

• Association for Computational Linguistics

- Note:
- Pages:

• 88–94

- Language:
- URL:
- [https://aclanthology.org/2026.arabicnlp-sharedtasks.10/](https://aclanthology.org/2026.arabicnlp-sharedtasks.10/)
- DOI:
- Cite (ACL):

• Musab Iskandar, Mohammed Al-Haig, Adel Sameer Alharbi, Khalid Alharbi, and Youssef Mohamed. 2026. AI Elites at NADI 2026 Shared Task: Selecting on Out-of-Domain Accuracy for Cross-Domain Spoken Arabic Dialect Identification . InProceedings of The Fourth Arabic Natural Language Processing Conference: Shared Tasks , pages 88–94, Budapest, Hungary. Association for Computational Linguistics.
• Cite (Informal):
• AI Elites at NADI 2026 Shared Task: Selecting on Out-of-Domain Accuracy for Cross-Domain Spoken Arabic Dialect Identification (Iskandar et al., ArabicNLP 2026)

- PDF:
- [https://aclanthology.org/2026.arabicnlp-sharedtasks.10.pdf](https://aclanthology.org/2026.arabicnlp-sharedtasks.10.pdf)