Fine-Tuning Whisper for Automatic Speech Recognition in Baniwa: A Preliminary Study
2026-08-26 • Computation and Language
Computation and Language
AI summaryⓘ
The authors studied how to use a popular speech recognition model called Whisper to understand Baniwa, a rare indigenous language spoken in parts of South America. They trained the model with a small collection of about half an hour of recorded words and short phrases. After training, the model could recognize Baniwa speech with some errors but showed it is possible to adapt large multilingual models to languages with very little data. This work provides a starting point for improving speech recognition in Baniwa and similar low-resource languages.
Automatic Speech RecognitionWhisper modelBaniwa languagelow-resource languagesmultilingual foundation modelsWord Error RateCharacter Error Ratefine-tuninglinguistic documentation
Authors
Leonardo Duart, Tiago Fonseca, Thiago Chacón
Abstract
Automatic Speech Recognition (ASR) technologies have achieved remarkable performance in recent years through the use of large multilingual foundation models. However, most advances remain concentrated on high-resource languages, while indigenous languages continue to suffer from a lack of speech resources and language technologies. This work presents a preliminary study on the adaptation of Whisper for Automatic Speech Recognition in Baniwa, an indigenous Arawakan language spoken in Brazil, Colombia, and Venezuela. The experiments were conducted using a corpus of 1,373 manually transcribed recordings obtained from a linguistic documentation project. The corpus contains approximately 0.54 hours of speech and consists primarily of isolated words and short elicited utterances. The Whisper Small model was fine-tuned using supervised learning and evaluated using Word Error Rate (WER) and Character Error Rate (CER). The best model achieved a WER of 37.5% and a CER of 7.45%, demonstrating that multilingual foundation models can be successfully adapted to extremely low-resource indigenous languages. The results establish an initial baseline for Baniwa Automatic Speech Recognition and provide a foundation for future research involving larger datasets, language-specific adaptation strategies, and post-processing techniques.