Simple and Effective Zero-shot Cross-lingual Phoneme Recognition
This addresses the challenge of speech recognition for unseen languages without labeled data, leveraging related language data, but it is incremental as it builds on existing transfer learning and pretraining techniques.
The paper tackles the problem of zero-shot cross-lingual phoneme recognition by fine-tuning a multilingually pretrained wav2vec 2.0 model and mapping phonemes using articulatory features, resulting in significant outperformance over prior work with simpler methods.
Recent progress in self-training, self-supervised pretraining and unsupervised learning enabled well performing speech recognition systems without any labeled data. However, in many cases there is labeled data available for related languages which is not utilized by these methods. This paper extends previous work on zero-shot cross-lingual transfer learning by fine-tuning a multilingually pretrained wav2vec 2.0 model to transcribe unseen languages. This is done by mapping phonemes of the training languages to the target language using articulatory features. Experiments show that this simple method significantly outperforms prior work which introduced task-specific architectures and used only part of a monolingually pretrained model.