Reduce, Reuse, Recycle: Is Perturbed Data better than Other Language augmentation for Low Resource Self-Supervised Speech Models

TL;DR

This paper investigates the effectiveness of various audio augmentation techniques for low-resource self-supervised speech models, finding combined synthetic augmentations outperform other methods in phoneme recognition tasks.

Contribution

It introduces a combined synthetic augmentation strategy for pre-training SSRL models in low-resource settings, outperforming accent and language transfer methods.

Findings

Combined noise and pitch augmentation outperforms other augmentation strategies.

Scaling augmented data can match the performance of target domain pre-training.

Synthetic augmentations are a viable alternative for resource-constrained languages.

Abstract

Self-supervised representation learning (SSRL) has demonstrated superior performance than supervised models for tasks including phoneme recognition. Training SSRL models poses a challenge for low-resource languages where sufficient pre-training data may not be available. A common approach is cross-lingual pre-training. Instead, we propose to use audio augmentation techniques, namely: pitch variation, noise addition, accented target language and other language speech to pre-train SSRL models in a low resource condition and evaluate phoneme recognition. Our comparisons found that a combined synthetic augmentations (noise/pitch) strategy outperformed accent and language knowledge transfer. Furthermore, we examined the scaling factor of augmented data to achieve equivalent performance to model pre-trained with target domain speech. Our findings suggest that for resource-constrained languages, combined augmentations can be a viable option than other augmentations.