Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

CROSS LINGUAL TRANSFER LEARNING FOR ZERO-RESOURCE DOMAIN ADAPTATION

Abstract: 

We propose a method for zero-resource domain adaptation of DNN acoustic models, for use in low-resource situations where the only in-language training data available may be poorly matched to the intended target domain. Our method uses a multi-lingual model in which several DNN layers are shared between languages. This architecture enables domain adaptation transforms learned for one well-resourced language to be applied to an entirely different low- resource language. First, to develop the technique we use English as a well-resourced language and take Spanish to mimic a low-resource language. Experiments in domain adaptation between the conversational telephone speech (CTS) domain and broadcast news (BN) domain demonstrate a 29% relative WER improvement on Spanish BN test data by using only English adaptation data. Second, we demonstrate the effectiveness of the method for low-resource languages with a poor match to the well-resourced language. Even in this scenario, the proposed method achieves relative WER improvements of 18-27% by using solely English data for domain adaptation. Compared to other related approaches based on multi-task and multi-condition training, the proposed method is able to better exploit well-resource language data for improved acoustic modelling of the low-resource target domain.

up
0 users have voted:

Paper Details

Authors:
Alberto Abad, Peter Bell, Andrea Carmantini, Steve Renals
Submitted On:
22 May 2020 - 8:32am
Short Link:
Type:
Presentation Slides
Event:
Presenter's Name:
Alberto Abad
Paper Code:
3775
Document Year:
2020
Cite

Document Files

ICASSP20_slides.pdf

(43)

Subscribe

[1] Alberto Abad, Peter Bell, Andrea Carmantini, Steve Renals, "CROSS LINGUAL TRANSFER LEARNING FOR ZERO-RESOURCE DOMAIN ADAPTATION", IEEE SigPort, 2020. [Online]. Available: http://sigport.org/5432. Accessed: Oct. 01, 2020.
@article{5432-20,
url = {http://sigport.org/5432},
author = {Alberto Abad; Peter Bell; Andrea Carmantini; Steve Renals },
publisher = {IEEE SigPort},
title = {CROSS LINGUAL TRANSFER LEARNING FOR ZERO-RESOURCE DOMAIN ADAPTATION},
year = {2020} }
TY - EJOUR
T1 - CROSS LINGUAL TRANSFER LEARNING FOR ZERO-RESOURCE DOMAIN ADAPTATION
AU - Alberto Abad; Peter Bell; Andrea Carmantini; Steve Renals
PY - 2020
PB - IEEE SigPort
UR - http://sigport.org/5432
ER -
Alberto Abad, Peter Bell, Andrea Carmantini, Steve Renals. (2020). CROSS LINGUAL TRANSFER LEARNING FOR ZERO-RESOURCE DOMAIN ADAPTATION. IEEE SigPort. http://sigport.org/5432
Alberto Abad, Peter Bell, Andrea Carmantini, Steve Renals, 2020. CROSS LINGUAL TRANSFER LEARNING FOR ZERO-RESOURCE DOMAIN ADAPTATION. Available at: http://sigport.org/5432.
Alberto Abad, Peter Bell, Andrea Carmantini, Steve Renals. (2020). "CROSS LINGUAL TRANSFER LEARNING FOR ZERO-RESOURCE DOMAIN ADAPTATION." Web.
1. Alberto Abad, Peter Bell, Andrea Carmantini, Steve Renals. CROSS LINGUAL TRANSFER LEARNING FOR ZERO-RESOURCE DOMAIN ADAPTATION [Internet]. IEEE SigPort; 2020. Available from : http://sigport.org/5432