English

SCaLa: Supervised Contrastive Learning for End-to-End Speech Recognition

Audio and Speech Processing 2022-06-22 v3 Machine Learning Sound

Abstract

End-to-end Automatic Speech Recognition (ASR) models are usually trained to optimize the loss of the whole token sequence, while neglecting explicit phonemic-granularity supervision. This could result in recognition errors due to similar-phoneme confusion or phoneme reduction. To alleviate this problem, we propose a novel framework based on Supervised Contrastive Learning (SCaLa) to enhance phonemic representation learning for end-to-end ASR systems. Specifically, we extend the self-supervised Masked Contrastive Predictive Coding (MCPC) to a fully-supervised setting, where the supervision is applied in the following way. First, SCaLa masks variable-length encoder features according to phoneme boundaries given phoneme forced-alignment extracted from a pre-trained acoustic model; it then predicts the masked features via contrastive learning. The forced-alignment can provide phoneme labels to mitigate the noise introduced by positive-negative pairs in self-supervised MCPC. Experiments on reading and spontaneous speech datasets show that our proposed approach achieves 2.8 and 1.4 points Character Error Rate (CER) absolute reductions compared to the baseline, respectively.

Keywords

Cite

@article{arxiv.2110.04187,
  title  = {SCaLa: Supervised Contrastive Learning for End-to-End Speech Recognition},
  author = {Li Fu and Xiaoxiao Li and Runyu Wang and Lu Fan and Zhengchen Zhang and Meng Chen and Youzheng Wu and Xiaodong He},
  journal= {arXiv preprint arXiv:2110.04187},
  year   = {2022}
}

Comments

INTERSPEECH 2022

R2 v1 2026-06-24T06:44:31.332Z