Kunnafonidilaw ka Cadeau: an ASR dataset of present-day Bambara
This work addresses the need for robust ASR in low-resource languages like Bambara, though it is incremental as it applies existing methods to new data.
The authors tackled the problem of automatic speech recognition (ASR) for Bambara, a predominantly oral language, by creating a 160-hour dataset from Malian radio archives that includes real-world challenges like code-switching and background noise. Finetuning models on this dataset reduced word error rates from 44.47% to 37.12% and from 36.07% to 32.33% on test sets, and it outperformed a model trained on cleaner data in human evaluation.
We present Kunkado, a 160-hour Bambara ASR dataset compiled from Malian radio archives to capture present-day spontaneous speech across a wide range of topics. It includes code-switching, disfluencies, background noise, and overlapping speakers that practical ASR systems encounter in real-world use. We finetuned Parakeet-based models on a 33.47-hour human-reviewed subset and apply pragmatic transcript normalization to reduce variability in number formatting, tags, and code-switching annotations. Evaluated on two real-world test sets, finetuning with Kunkado reduces WER from 44.47\% to 37.12\% on one and from 36.07\% to 32.33\% on the other. In human evaluation, the resulting model also outperforms a comparable system with the same architecture trained on 98 hours of cleaner, less realistic speech. We release the data and models to support robust ASR for predominantly oral languages.