Self-supervised Rewiring of Pre-trained Speech Encoders: Towards Faster Fine-tuning with Less Labels in Speech Processing

TL;DR

This paper introduces a self-supervised rewiring method for pre-trained speech encoders that improves their representation space, leading to faster and more effective fine-tuning, especially with limited labeled data.

Contribution

The authors propose a label-free, contrastive self-supervised rewiring technique that enhances speech encoder representations and accelerates downstream task fine-tuning.

Findings

Improved isotropy in the representation space of wav2vec 2.

Significant speedup in fine-tuning convergence across 6 speech tasks.

Consistent performance gains in low-resource scenarios.

Abstract

Pre-trained speech Transformers have facilitated great success across various speech processing tasks. However, fine-tuning these encoders for downstream tasks require sufficiently large training data to converge or to achieve state-of-the-art. In text domain this has been partly attributed to sub-optimality of the representation space in pre-trained Transformers. In this work, we take a sober look into pre-trained speech encoders and rewire their representation space without requiring any task-specific labels. Our method utilises neutrally synthesised version of audio inputs along with frame masking to construct positive pairs for contrastive self-supervised learning. When used for augmenting the wav2vec 2 encoder, we observe consistent improvement of isotropy in the representation space. Our experiments on 6 speech processing tasks, exhibit a significant convergence speedup during task fine-tuning as well as consistent task improvement, specially in low-resource settings.