vq-wav2vec: Self-Supervised Learning of Discrete Speech Representations

Alexei Baevski, Steffen Schneider, Michael Auli

Abstract

We propose vq-wav2vec to learn discrete representations of audio segments through a wav2vec-style self-supervised context prediction task. The algorithm uses either a gumbel softmax or online k-means clustering to quantize the dense representations. Discretization enables the direct application of algorithms from the NLP community which require discrete inputs. Experiments show that BERT pre-training achieves a new state of the art on TIMIT phoneme classification and WSJ speech recognition.
BibTeX
@inproceedings{baevski2020vq,
  author = {Alexei Baevski and Steffen Schneider and Michael Auli},
  title = {{vq-wav2vec: Self-Supervised Learning of Discrete Speech Representations}},
  booktitle = {International Conference on Learning Representations},
  year = {2020},
  url = {https://openreview.net/forum?id=75lAdqEmCvk},
}
Figure