This paper presents a novel acoustic modeling technique of large vocabulary automatic speech recognition for under-resourced languages by leveraging well-trained acoustic models of other languages (called source languages). The idea is to use source language acoustic model to score the acoustic features of the target language, and then map these scores to the posteriors of the target phones using a classifier. The target phone posteriors are then used for decoding in the usual way of hybrid acoustic modeling. The motivation of such a strategy is that human languages usually share similar phone sets and hence it may be easier to predict the target phone posteriors from the scores generated by source language acoustic models than to train from scratch an under-resourced language acoustic model. The proposed method is evaluated using on the Aurora-4 task with less than 1 hour of training data. Two types of source language acoustic models are considered, i.e. hybrid HMM/MLP and conventional HMM/GMM models. In addition, we also use triphone tied states in the mapping. Our experimental results show that by leveraging well trained Malay and Hungarian acoustic models, we achieved 9.0% word error rate (WER) given 55 minutes of English training data. This is close to the WER of 7.9% obtained by using the full 15 hours of training data and much better than the WER of 14.4% obtained by conventional acoustic modeling techniques with the same 55 minutes of training data.
Van Hai DO
Nanyang Technological University
Xiong XIAO
Nanyang Technological University
Eng Siong CHNG
Nanyang Technological University
Haizhou LI
Nanyang Technological University,Institute for Infocomm Research
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Van Hai DO, Xiong XIAO, Eng Siong CHNG, Haizhou LI, "Cross-Lingual Phone Mapping for Large Vocabulary Speech Recognition of Under-Resourced Languages" in IEICE TRANSACTIONS on Information,
vol. E97-D, no. 2, pp. 285-295, February 2014, doi: 10.1587/transinf.E97.D.285.
Abstract: This paper presents a novel acoustic modeling technique of large vocabulary automatic speech recognition for under-resourced languages by leveraging well-trained acoustic models of other languages (called source languages). The idea is to use source language acoustic model to score the acoustic features of the target language, and then map these scores to the posteriors of the target phones using a classifier. The target phone posteriors are then used for decoding in the usual way of hybrid acoustic modeling. The motivation of such a strategy is that human languages usually share similar phone sets and hence it may be easier to predict the target phone posteriors from the scores generated by source language acoustic models than to train from scratch an under-resourced language acoustic model. The proposed method is evaluated using on the Aurora-4 task with less than 1 hour of training data. Two types of source language acoustic models are considered, i.e. hybrid HMM/MLP and conventional HMM/GMM models. In addition, we also use triphone tied states in the mapping. Our experimental results show that by leveraging well trained Malay and Hungarian acoustic models, we achieved 9.0% word error rate (WER) given 55 minutes of English training data. This is close to the WER of 7.9% obtained by using the full 15 hours of training data and much better than the WER of 14.4% obtained by conventional acoustic modeling techniques with the same 55 minutes of training data.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E97.D.285/_p
Copy
@ARTICLE{e97-d_2_285,
author={Van Hai DO, Xiong XIAO, Eng Siong CHNG, Haizhou LI, },
journal={IEICE TRANSACTIONS on Information},
title={Cross-Lingual Phone Mapping for Large Vocabulary Speech Recognition of Under-Resourced Languages},
year={2014},
volume={E97-D},
number={2},
pages={285-295},
abstract={This paper presents a novel acoustic modeling technique of large vocabulary automatic speech recognition for under-resourced languages by leveraging well-trained acoustic models of other languages (called source languages). The idea is to use source language acoustic model to score the acoustic features of the target language, and then map these scores to the posteriors of the target phones using a classifier. The target phone posteriors are then used for decoding in the usual way of hybrid acoustic modeling. The motivation of such a strategy is that human languages usually share similar phone sets and hence it may be easier to predict the target phone posteriors from the scores generated by source language acoustic models than to train from scratch an under-resourced language acoustic model. The proposed method is evaluated using on the Aurora-4 task with less than 1 hour of training data. Two types of source language acoustic models are considered, i.e. hybrid HMM/MLP and conventional HMM/GMM models. In addition, we also use triphone tied states in the mapping. Our experimental results show that by leveraging well trained Malay and Hungarian acoustic models, we achieved 9.0% word error rate (WER) given 55 minutes of English training data. This is close to the WER of 7.9% obtained by using the full 15 hours of training data and much better than the WER of 14.4% obtained by conventional acoustic modeling techniques with the same 55 minutes of training data.},
keywords={},
doi={10.1587/transinf.E97.D.285},
ISSN={1745-1361},
month={February},}
Copy
TY - JOUR
TI - Cross-Lingual Phone Mapping for Large Vocabulary Speech Recognition of Under-Resourced Languages
T2 - IEICE TRANSACTIONS on Information
SP - 285
EP - 295
AU - Van Hai DO
AU - Xiong XIAO
AU - Eng Siong CHNG
AU - Haizhou LI
PY - 2014
DO - 10.1587/transinf.E97.D.285
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E97-D
IS - 2
JA - IEICE TRANSACTIONS on Information
Y1 - February 2014
AB - This paper presents a novel acoustic modeling technique of large vocabulary automatic speech recognition for under-resourced languages by leveraging well-trained acoustic models of other languages (called source languages). The idea is to use source language acoustic model to score the acoustic features of the target language, and then map these scores to the posteriors of the target phones using a classifier. The target phone posteriors are then used for decoding in the usual way of hybrid acoustic modeling. The motivation of such a strategy is that human languages usually share similar phone sets and hence it may be easier to predict the target phone posteriors from the scores generated by source language acoustic models than to train from scratch an under-resourced language acoustic model. The proposed method is evaluated using on the Aurora-4 task with less than 1 hour of training data. Two types of source language acoustic models are considered, i.e. hybrid HMM/MLP and conventional HMM/GMM models. In addition, we also use triphone tied states in the mapping. Our experimental results show that by leveraging well trained Malay and Hungarian acoustic models, we achieved 9.0% word error rate (WER) given 55 minutes of English training data. This is close to the WER of 7.9% obtained by using the full 15 hours of training data and much better than the WER of 14.4% obtained by conventional acoustic modeling techniques with the same 55 minutes of training data.
ER -