English

MixSpeech: Data Augmentation for Low-resource Automatic Speech Recognition

Computation and Language 2021-02-26 v1 Sound Audio and Speech Processing

Abstract

In this paper, we propose MixSpeech, a simple yet effective data augmentation method based on mixup for automatic speech recognition (ASR). MixSpeech trains an ASR model by taking a weighted combination of two different speech features (e.g., mel-spectrograms or MFCC) as the input, and recognizing both text sequences, where the two recognition losses use the same combination weight. We apply MixSpeech on two popular end-to-end speech recognition models including LAS (Listen, Attend and Spell) and Transformer, and conduct experiments on several low-resource datasets including TIMIT, WSJ, and HKUST. Experimental results show that MixSpeech achieves better accuracy than the baseline models without data augmentation, and outperforms a strong data augmentation method SpecAugment on these recognition tasks. Specifically, MixSpeech outperforms SpecAugment with a relative PER improvement of 10.6%\% on TIMIT dataset, and achieves a strong WER of 4.7%\% on WSJ dataset.

Keywords

Cite

@article{arxiv.2102.12664,
  title  = {MixSpeech: Data Augmentation for Low-resource Automatic Speech Recognition},
  author = {Linghui Meng and Jin Xu and Xu Tan and Jindong Wang and Tao Qin and Bo Xu},
  journal= {arXiv preprint arXiv:2102.12664},
  year   = {2021}
}

Comments

To appear at ICASSP 2021

R2 v1 2026-06-23T23:29:39.108Z