aiqbench analysis
Cardamom introduces a critical dataset of 40 hours covering 21 micro-dialects to address the persistent gap in Automatic Speech Recognition (ASR) for fine-grained Arabic varieties. By benchmarking multilingual models, the study demonstrates that adaptation on this localized data significantly reduces word error rates compared to zero-shot performance. This resource is essential for advancing ASR systems beyond broad country-level labels to capture sub-national linguistic diversity.
CARDAMOM: A Micro-Dialectal Arabic Speech Dataset for ASR
arXiv:2609.34481v1 Announce Type: new Abstract: We present Cardamom, a micro-dialectal Arabic speech dataset designed to support fine-grained evaluation and adaptation of automatic speech recognition (ASR) systems. Community-curated by native speakers familiar with the represented varieties, Cardamom contains approximately 40 hours of transcribed YouTube speech spanning 21 micro-dialects across Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia. Each segment is annotated with one or more operational micro-dialect labels, code-switching information, and utterance-level perceived gender, enabling analysis of sub-country variation that is obscured by conventional country-level labels. We describe the collection and annotation process, motivate the micro-dialect inventory linguistically, and benchmark four multilingual ASR systems in zero-shot and adapted settings. The strongest zero-shot system obtains 43.47% aggregate WER, with particularly high error rates on Mauritanian and Lebanese varieties; adaptation on Cardamom reduces its WER to 35.21%. Audio-based identification experiments further show that the annotations provide a learnable prediction target, with a dedicated classifier reaching 85.57% accuracy on 21-way micro-dialect identification. Cardamom provides a resource for studying localized dialectal variation and developing Arabic speech systems with broader regional coverage.
Authors:Bashar Talafha, Samar M. Magdy, Aisha Alansari, Alaa Alkhawaldeh, Abdurrahman Juma, Sharaf Makahleh, Nour Gamal, Omar Attia, Hanaa Kurdi, Najwa Rizk, Maysa Anaya, Hessah Altimyat, Layal Alhazmi, Shumukh Alotaibi, Hajar Alhadaris, Rayan Alomari, Rahaf Almalaq, Malak Alkhorasani, Sara alghamdi, Rahaf Alshamrani, Nsrin Ashraf, Ibrahim Jaradat, Nada Qardahji, Yasmin Zaraket, Elmoukhtar Brahim, Sidi Ebeidy, Oumoulmouminin Mahmoud, Yahjeb Bouha Khatraty, Meya Haroune, Mohammad Ghaddar, Mohamad Eldirany, Rashed Alamoush, Tala Chhaytle, Nuha Albadi, Yahya El Hadj, Hamzah Luqman, Fadi A. Zaraket, Mustafa Jarrar, Muhammad Abdul-Mageed
Abstract:We present Cardamom, a micro-dialectal Arabic speech dataset designed to support fine-grained evaluation and adaptation of automatic speech recognition (ASR) systems. Community-curated by native speakers familiar with the represented varieties, Cardamom contains approximately 40 hours of transcribed YouTube speech spanning 21 micro-dialects across Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia. Each segment is annotated with one or more operational micro-dialect labels, code-switching information, and utterance-level perceived gender, enabling analysis of sub-country variation that is obscured by conventional country-level labels. We describe the collection and annotation process, motivate the micro-dialect inventory linguistically, and benchmark four multilingual ASR systems in zero-shot and adapted settings. The strongest zero-shot system obtains 43.47% aggregate WER, with particularly high error rates on Mauritanian and Lebanese varieties; adaptation on Cardamom reduces its WER to 35.21%. Audio-based identification experiments further show that the annotations provide a learnable prediction target, with a dedicated classifier reaching 85.57% accuracy on 21-way micro-dialect identification. Cardamom provides a resource for studying localized dialectal variation and developing Arabic speech systems with broader regional coverage.
Computation and Language (cs.CL)
arXiv:2609.34481 [cs.CL]
(or arXiv:2609.34481v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2609.34481 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 272 words
Related coverage
ArFake: A Robust Framework for Multi-Dialect Arabic Speech Spoofing Detection Benchmark
arXiv · Jun 17, 2026
Bulbul: A Dataset for Dialectal Arabic Speech Recognition
arXiv · Aug 25, 2026
Fine-Tuning NVIDIA Nemotron for Saudi Arabic Dialects, with a Path to Other Languages
NVIDIA · Oct 1, 2026
Towards Robust Arabic Speech Emotion Recognition with Deep Learning
arXiv · Jun 10, 2026