IEICE TRANSACTIONS on Information
Effective Prediction of Errors by Non-native Speakers Using Decision Tree for Speech Recognition-Based CALL System
Summary :
CALL (Computer Assisted Language Learning) systems using ASR (Automatic Speech Recognition) for second language learning have received increasing interest recently. However, it still remains a challenge to achieve high speech recognition performance, including accurate detection of erroneous utterances by non-native speakers. Conventionally, possible error patterns, based on linguistic knowledge, are added to the lexicon and language model, or the ASR grammar network. However, this approach easily falls in the trade-off of coverage of errors and the increase of perplexity. To solve the problem, we propose a method based on a decision tree to learn effective prediction of errors made by non-native speakers. An experimental evaluation with a number of foreign students learning Japanese shows that the proposed method can effectively generate an ASR grammar network, given a target sentence, to achieve both better coverage of errors and smaller perplexity, resulting in significant improvement in ASR accuracy.
- Publication
- IEICE TRANSACTIONS on Information Vol.E92-D No.12 pp.2462-2468
- Publication Date
- 2009/12/01
- Publicized
- Online ISSN
- 1745-1361
- DOI
- 10.1587/transinf.E92.D.2462
- Type of Manuscript
- PAPER
- Category
- Speech and Hearing
Authors
Keyword
Latest Issue
Copyrights notice of machine-translated contents
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.
Cite this
Copy
Hongcui WANG, Tatsuya KAWAHARA, "Effective Prediction of Errors by Non-native Speakers Using Decision Tree for Speech Recognition-Based CALL System" in IEICE TRANSACTIONS on Information,
vol. E92-D, no. 12, pp. 2462-2468, December 2009, doi: 10.1587/transinf.E92.D.2462.
Abstract: CALL (Computer Assisted Language Learning) systems using ASR (Automatic Speech Recognition) for second language learning have received increasing interest recently. However, it still remains a challenge to achieve high speech recognition performance, including accurate detection of erroneous utterances by non-native speakers. Conventionally, possible error patterns, based on linguistic knowledge, are added to the lexicon and language model, or the ASR grammar network. However, this approach easily falls in the trade-off of coverage of errors and the increase of perplexity. To solve the problem, we propose a method based on a decision tree to learn effective prediction of errors made by non-native speakers. An experimental evaluation with a number of foreign students learning Japanese shows that the proposed method can effectively generate an ASR grammar network, given a target sentence, to achieve both better coverage of errors and smaller perplexity, resulting in significant improvement in ASR accuracy.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E92.D.2462/_p
Copy
@ARTICLE{e92-d_12_2462,
author={Hongcui WANG, Tatsuya KAWAHARA, },
journal={IEICE TRANSACTIONS on Information},
title={Effective Prediction of Errors by Non-native Speakers Using Decision Tree for Speech Recognition-Based CALL System},
year={2009},
volume={E92-D},
number={12},
pages={2462-2468},
abstract={CALL (Computer Assisted Language Learning) systems using ASR (Automatic Speech Recognition) for second language learning have received increasing interest recently. However, it still remains a challenge to achieve high speech recognition performance, including accurate detection of erroneous utterances by non-native speakers. Conventionally, possible error patterns, based on linguistic knowledge, are added to the lexicon and language model, or the ASR grammar network. However, this approach easily falls in the trade-off of coverage of errors and the increase of perplexity. To solve the problem, we propose a method based on a decision tree to learn effective prediction of errors made by non-native speakers. An experimental evaluation with a number of foreign students learning Japanese shows that the proposed method can effectively generate an ASR grammar network, given a target sentence, to achieve both better coverage of errors and smaller perplexity, resulting in significant improvement in ASR accuracy.},
keywords={},
doi={10.1587/transinf.E92.D.2462},
ISSN={1745-1361},
month={December},}
Copy
TY - JOUR
TI - Effective Prediction of Errors by Non-native Speakers Using Decision Tree for Speech Recognition-Based CALL System
T2 - IEICE TRANSACTIONS on Information
SP - 2462
EP - 2468
AU - Hongcui WANG
AU - Tatsuya KAWAHARA
PY - 2009
DO - 10.1587/transinf.E92.D.2462
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E92-D
IS - 12
JA - IEICE TRANSACTIONS on Information
Y1 - December 2009
AB - CALL (Computer Assisted Language Learning) systems using ASR (Automatic Speech Recognition) for second language learning have received increasing interest recently. However, it still remains a challenge to achieve high speech recognition performance, including accurate detection of erroneous utterances by non-native speakers. Conventionally, possible error patterns, based on linguistic knowledge, are added to the lexicon and language model, or the ASR grammar network. However, this approach easily falls in the trade-off of coverage of errors and the increase of perplexity. To solve the problem, we propose a method based on a decision tree to learn effective prediction of errors made by non-native speakers. An experimental evaluation with a number of foreign students learning Japanese shows that the proposed method can effectively generate an ASR grammar network, given a target sentence, to achieve both better coverage of errors and smaller perplexity, resulting in significant improvement in ASR accuracy.
ER -
English
