Cardamom is presented, a micro-dialectal Arabic speech dataset designed to support fine-grained evaluation and adaptation of automatic speech recognition (ASR) systems, and the collection and annotation process is described, the micro-dialect inventory linguistically is motivated, and four multilingual ASR systems in zero-shot and adapted settings are benchmarked.
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 highlight the difficulty of fine-grained micro-dialect recognition: across three training seeds, a dedicated classifier obtains 46.85% accuracy and 31.95% F1. Cardamom provides a resource for studying localized dialectal variation and developing Arabic speech systems with broader regional coverage.
ALMIEYAR is introduced, a culturally grounded ASR benchmark covering 17 Arabic dialects across six families, built entirely from newly recorded speech unseen by existing models, including the first published benchmark for Ahwazi Arabic.
Omid Ghahroodi, Anas Madkoor, Dima Faris Alsaudi et al.· arXiv.org· 0 citations
Vietnamese speech research is constrained by resources that isolate automatic speech recognition from speaker, dialect, code-switching, and deepfake analysis. We introduce VietPrism, an open, multi-domain corpus that brings these dimensions together at scale: 993.4 hours and 403,941 bona fide utterances from 1,262 veri...
This work introduces the first multitask corpus of Tunisian dialect manually aligned with its equivalents in modern standard Arabic, and designs three variants: a baseline multitask classifier, a cross-attention model aligning Tunisian dialect and modern standard Arabic representations, and a context-aware cross-attent...
Roua Torjmen, K. Haddar· International Conference on...· 0 citations
The findings support a diagnostic rather than leaderboard interpretation: the apparent benefit of Arabic XTTS adaptation depends materially on the evaluation metric, checkpoint, and evaluator.
Results show that out-of-domain generalization remains a major bottleneck and highlight the effectiveness of recent Arabic-specialized speech models, multimodal dialect identification approaches, and ensemble methods.
Peter Sullivan, Bashar Talafha, Ahmed Ashraf et al.· 4 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 19, 2026
An MIT and Dana-Farber Institute-led team used large language models to examine nearly 4,000 studies and hundreds of code repositories, revealing gaps in the information needed to scrutinize and reuse published research.
By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.