Self-Train Before You Transcribe
–arXiv.org Artificial Intelligence
When there is a mismatch between the training and test domains, current speech recognition systems show significant performance degradation. Self-training methods, such as noisy student teacher training, can help address this and enable the adaptation of models under such domain shifts. However, self-training typically requires a collection of unlabelled target domain data. For settings where this is not practical, we investigate the benefit of performing noisy student teacher training on recordings in the test set as a test-time adaptation approach. Similarly to the dynamic evaluation approach in language modelling, this enables the transfer of information across utterance boundaries and functions as a method of domain adaptation. A range of in-domain and out-of-domain datasets are used for experiments demonstrating large relative gains of up to 32.2%. Interestingly, our method showed larger gains than the typical self-training setup that utilises separate adaptation data.
arXiv.org Artificial Intelligence
Jun-17-2024
- Country:
- Europe
- Germany (0.14)
- United Kingdom (0.14)
- Europe
- Genre:
- Research Report > New Finding (0.46)
- Industry:
- Education > Teacher Education (0.44)
- Technology: