Metraipolis › archive › dc_8f59ca574b29
How Architecture, Typology, and Adaptation Strategies Shape ASR for Low-Resource African Languages: A Controlled Empirical Study — arXiv:2610.39004
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.39004
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T06:16:56.078Z
- SHA-256 of the archived text
4de9436f676e0841d523c5b9444fe81f18540a2a25f9f3dd674c682f514e0c59- Stable URL of this page
https://metraipolis.com/documento/dc_8f59ca574b29
How to cite
How Architecture, Typology, and Adaptation Strategies Shape ASR for Low-Resource African Languages: A Controlled Empirical Study — arXiv:2610.39004 — arXiv (en), https://arxiv.org/abs/2610.39004 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_8f59ca574b29 (Metraipolis archive dc_8f59ca574b29, sha256 4de9436f676e0841d523c5b9444fe81f18540a2a25f9f3dd674c682f514e0c59)
BibTeX
@misc{dc8f59ca574b29,
title = {How Architecture, Typology, and Adaptation Strategies Shape ASR for Low-Resource African Languages: A Controlled Empirical Study — arXiv:2610.39004},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.39004},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_8f59ca574b29; sha256 4de9436f676e0841d523c5b9444fe81f18540a2a25f9f3dd674c682f514e0c59},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.39004. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Despite recent advancements in ASR, identifying effective architectures, fine-tuning approaches, and data adaptation strategies for low-resource languages remains challenging. We present a systematic comparison across five low-resource African languages and Swahili as a mid-resource contrast, fine-tuning three models from distinct architectures: Phi-4-multimodal-instruct, Whisper-large-v3-turbo, and MMS-1B-all. We analyze data sufficiency, monolingual versus multilingual adaptation, domain fine-tuning, and language-adapter transfer. Phi-4 achieves the best overall accuracy, MMS shows comparable CER, while Whisper exhibits high CER, consistent with known hallucinations and repetition loops. Performance saturates at different data scales, offering practical data-efficiency targets, while typological complexity consistently shapes error rates. Domain fine-tuning benefits Phi-4 and Whisper but degrades MMS, revealing limitations of adapter-based adaptation. Proxy language adapters outperform training MMS adapters from scratch. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno). Categorie arXiv: Computation and Language (cs.CL) ; Sound (cs.SD); Audio and Speech Processing (eess.AS). Pagina del lavoro: https://arxiv.org/abs/2610.39004 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-31T06:07:23.347Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).