When Is TTS Augmentation Through a Pivot Language Useful?
This addresses the challenge of limited transcribed audio data for low-resource language ASR, though it is incremental as it builds on existing TTS and data augmentation techniques.
The paper tackles the problem of developing Automatic Speech Recognition (ASR) for low-resource languages by proposing to use synthetic audio generated via text-to-speech (TTS) systems from a pivot language, finding that this approach can improve ASR performance with concrete gains of 64.5% and 45.0% character error reduction rate for Guaraní and Suba, respectively.
Developing Automatic Speech Recognition (ASR) for low-resource languages is a challenge due to the small amount of transcribed audio data. For many such languages, audio and text are available separately, but not audio with transcriptions. Using text, speech can be synthetically produced via text-to-speech (TTS) systems. However, many low-resource languages do not have quality TTS systems either. We propose an alternative: produce synthetic audio by running text from the target language through a trained TTS system for a higher-resource pivot language. We investigate when and how this technique is most effective in low-resource settings. In our experiments, using several thousand synthetic TTS text-speech pairs and duplicating authentic data to balance yields optimal results. Our findings suggest that searching over a set of candidate pivot languages can lead to marginal improvements and that, surprisingly, ASR performance can by harmed by increases in measured TTS quality. Application of these findings improves ASR by 64.5\% and 45.0\% character error reduction rate (CERR) respectively for two low-resource languages: Guaraní and Suba.