{
  "id": 740,
  "url": "https://arxiv.org/abs/2606.22274v1",
  "title": "From Speech to Text Corpora: Evaluating ASR-Based Data Acquisition for Low-Resource Fongbe and Hausa",
  "summary": "Low-resource African languages lack text corpora needed for language model training. We investigate whether ASR pipelines can extend text resources for two typologically distinct West African languages: Fongbe (tonal, diacritic-rich) and Hausa (non-tonal). We fine-tune MMS-300M on a curated 12.3-hour Fongbe dataset, achieving 9.48% WER on the ALFFA benchmark - a 78% relative reduction from the prior 44.04% baseline - while preserving tonal diacritics critical to the language. For Hausa, we apply",
  "authors": "Mahounan Pericles Adjovi, Victor Olufemi, Roald Eiselen, Prasenjit Mitra",
  "category": "research",
  "topics": "finance-investment",
  "orgs": null,
  "regions": "africa",
  "published_at": "2026-06-20T23:51:55.000Z",
  "fetched_at": "2026-07-14T14:14:46.034Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/740",
  "original_url": "https://arxiv.org/abs/2606.22274v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}