ICASSP 2023accepted0 citations
Alignment Entropy Regularization
Ehsan Variani, Ke Wu, David Rybach, Cyril Allauzen, Michael Riley
Abstract
Existing training criteria in automatic speech recognition (ASR) permit the model to freely explore more than one time alignments between the feature and label sequences. In this paper, we use entropy to measure a model’s uncertainty, i.e. how it chooses to distribute the probability mass over the set of allowed alignments. Furthermore, we evaluate the effect of entropy regularization in encouraging the model to distribute the probability mass only on a smaller subset of allowed alignments. Experiments show that entropy regularization enables a much simpler decoding method without sacrificing word error rate, and provides better time alignment quality.
BibTeX
@inproceedings{icassp2023_alignmententropy,
title = {Alignment Entropy Regularization},
author = {Ehsan Variani and Ke Wu and David Rybach and Cyril Allauzen and Michael Riley},
booktitle = {ICASSP 2023},
year = {2023}
}