Masked Audio Text Encoders are Effective Multi-Modal Rescorers
Jinglun Cai, Monica Sunkara, Xilai Li, Anshu Bhatia, Xiao Pan, Sravan Bodapati
Abstract
Masked Language Models (MLMs) have proven to be effective for second-pass rescoring in Automatic Speech Recognition (ASR) systems. In this work, we propose Masked Audio Text Encoder (MATE), a multi-modal masked language model rescorer which incorporates acoustic representations into the input space of MLM. We adopt contrastive learning for effectively aligning the modalities by learning shared representations. We show that using a multi-modal rescorer is beneficial for domain generalization of the ASR system when target domain data is unavailable. MATE reduces word error rate (WER) by 4%-16% on in-domain, and 3%-7% on out-of-domain datasets, over the text-only baseline. Additionally, with very limited amount of training data (0.8 hours) MATE achieves a WER reduction of 8%-23% over the first-pass baseline.
BibTeX
@inproceedings{cai-etal-2023-masked,
title = "Masked Audio Text Encoders are Effective Multi-Modal Rescorers",
author = "Cai, Jinglun and
Sunkara, Monica and
Li, Xilai and
Bhatia, Anshu and
Pan, Xiao and
Bodapati, Sravan",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2023",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-acl.682/",
doi = "10.18653/v1/2023.findings-acl.682",
pages = "10718--10730"
}