Publication Details
Text Augmentation for Language Models in High Error Recognition Scenario
data augmentation, error simulation, language modeling, automatic speech
recognition
In this paper, we explore several data augmentation strategies for training of
language models for speech recognition. We compare augmentation based on global
error statistics with one based on unigram statistics of ASR errors and with
labelsmoothing and its sampled variant. Additionally, we investigate the
stability and the predictive power of perplexity estimated on augmented data.
Despite being trivial, augmentation driven by global substitution, deletion and
insertion rates achieves the best rescoring results. On the other hand, even
though the associated perplexity measure is stable, it gives no better prediction
of the final error rate than the vanilla one. Our best augmentation scheme
increases the WER improvement from second-pass rescoring from 1.1% to 1.9%
absolute on the CHiMe-6 challenge.
@inproceedings{BUT175841,
author="Karel {Beneš} and Lukáš {Burget}",
title="Text Augmentation for Language Models in High Error Recognition Scenario",
booktitle="Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH",
year="2021",
journal="Proceedings of Interspeech",
volume="2021",
number="8",
pages="1872--1876",
publisher="International Speech Communication Association",
address="Brno",
doi="10.21437/Interspeech.2021-627",
issn="1990-9772",
url="https://www.isca-speech.org/archive/interspeech_2021/benes21_interspeech.html"
}