Resource-Efficient Fine-Tuning Strategies for Automatic MOS Prediction in Text-to-Speech for Low-Resource Languages
Abstract
We train a MOS prediction model based on wav2vec 2.0 using the open-access data sets BVCC and SOMOS. Our test with neural TTS data in the low-resource language (LRL) West Frisian shows that pre-training on BVCC before fine-tuning on SOMOS leads to the best accuracy for both fine-tuned and zero-shot prediction. Further fine-tuning experiments show that using more than 30 percent of the total data does not lead to significant improvements. In addition, fine-tuning with data from a single listener shows promising system-level accuracy, supporting the viability of one-participant pilot tests. These findings can all assist the resource-conscious development of TTS for LRLs by progressing towards better zero-shot MOS prediction and informing the design of listening tests, especially in early-stage evaluation.
Type
Publication
Proceedings of INTERSPEECH 2023
This paper was accepted at INTERSPEECH 2023. It explores resource-efficient strategies for automatic Mean Opinion Score (MOS) prediction in text-to-speech systems for low-resource languages, with a particular focus on West Frisian. The research demonstrates that pre-training on larger datasets before fine-tuning on smaller, targeted datasets yields the best accuracy for MOS prediction, and that using as little as 30% of available data can produce effective results.
