Publication Details
Ask2Mask: Guided Data Selection for Masked Speech Modeling
Rosenberg Andrew
Ramabhadran Bhuvana
Zhang Yu
Moreno Pedro
Guided Data Selection, Masked Speech Modeling
Masked speech modeling (MSM) methods such as wav2vec2 or w2v-BERT learn
representations over speech frames which are randomlymaskedwithin an utterance.
While thesemethods improve performance of Automatic Speech Recognition (ASR)
systems, they have one major limitation. They treat all unsupervised speech
samples with equal weight, which hinders learning as not all samples have
relevant information to learn meaningful representations. In this work, we
address this limitation. We propose ask2mask (ATM), a novel approach to focus on
specific samples during MSM pre-training. ATM employs an external ASR model or
scorer to weight unsupervised input samples in two different ways: 1)
A fine-grained data selection is performed by masking over the highly confident
input frames as chosen by the scorer. This allows themodel to learnmeaningful
representations. 2) ATM is further extended to focus at utterance-level by
weighting the final MSM loss with the utterance-level confidence score. We
conduct fine-tuning experiments on two well-benchmarked corpora: LibriSpeech
(matching the pre-training data) and Commonvoice, TED-LIUM, AMI and CHiME-6 (not
matching the pre-training data). The results substantiate the efficacy of ATM on
significantly improving the recognition performance under mismatched conditions
(up to 11.6% relative over published results and upto 4.46% relative over our
internal baseline) while still yielding modest improvements under matched
conditions.
@article{BUT182529,
author="Murali Karthick {Baskar} and Andrew {Rosenberg} and Bhuvana {Ramabhadran} and Yu {Zhang} and Pedro {Moreno}",
title="Ask2Mask: Guided Data Selection for Masked Speech Modeling",
journal="IEEE J-STSP",
year="2022",
volume="16",
number="6",
pages="1357--1366",
doi="10.1109/JSTSP.2022.3186162",
issn="1932-4553",
url="https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9806175"
}