Token-wise Training for Attention Based End-to-end Speech Recognition

Peidong Wang, Jia Cui, Chao Weng, Dong Yu

Abstract

In attention based end-to-end (A-E2E) speech recognition systems, the dependency between output tokens is typically formulated as an input-output mapping in decoder. Due to such dependency, decoding errors can easily propagate along output sequence. In this paper, we propose a token-wise training (TWT) method for A-E2E models. The new method is flexible and can be combined with a variety of loss functions. Applying TWT to multiple hypotheses, we propose a novel TWT in beam (TWTiB) training scheme. Trained on the benchmark Switchboard (SWBD) 300h corpus, TWTiB outperforms the previous best training scheme on the SWBD evaluation subset.
BibTeX
@inproceedings{wang2019token,
  author = {Peidong Wang and Jia Cui and Chao Weng and Dong Yu},
  title = {{Token-wise Training for Attention Based End-to-end Speech Recognition}},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing},
  year = {2019},
  doi = {10.1109/icassp.2019.8683302},
}

Loading related content...

Figure