Exploring Non-Autoregressive End-To-End Neural Modeling For English Mispronunciation Detection And Diagnosis
End-to-end (E2E) neural modeling has emerged as one predominant school of thought to develop computer-assisted language training (CAPT) systems, showing competitive performance to conventional pronunciation-scoring based methods. However, current E2E neural methods for CAPT are faced with at least t...
Main Authors: | , , , , |
---|---|
Format: | Article in Journal/Newspaper |
Language: | unknown |
Published: |
arXiv
2021
|
Subjects: | |
Online Access: | https://dx.doi.org/10.48550/arxiv.2111.00844 https://arxiv.org/abs/2111.00844 |
id |
ftdatacite:10.48550/arxiv.2111.00844 |
---|---|
record_format |
openpolar |
spelling |
ftdatacite:10.48550/arxiv.2111.00844 2023-05-15T15:09:30+02:00 Exploring Non-Autoregressive End-To-End Neural Modeling For English Mispronunciation Detection And Diagnosis Wang, Hsin-Wei Yan, Bi-Cheng Chiu, Hsuan-Sheng Hsu, Yung-Chang Chen, Berlin 2021 https://dx.doi.org/10.48550/arxiv.2111.00844 https://arxiv.org/abs/2111.00844 unknown arXiv arXiv.org perpetual, non-exclusive license http://arxiv.org/licenses/nonexclusive-distrib/1.0/ Computation and Language cs.CL Multimedia cs.MM FOS Computer and information sciences Article CreativeWork article Preprint 2021 ftdatacite https://doi.org/10.48550/arxiv.2111.00844 2022-03-10T14:48:18Z End-to-end (E2E) neural modeling has emerged as one predominant school of thought to develop computer-assisted language training (CAPT) systems, showing competitive performance to conventional pronunciation-scoring based methods. However, current E2E neural methods for CAPT are faced with at least two pivotal challenges. On one hand, most of the E2E methods operate in an autoregressive manner with left-to-right beam search to dictate the pronunciations of an L2 learners. This however leads to very slow inference speed, which inevitably hinders their practical use. On the other hand, E2E neural methods are normally data greedy and meanwhile an insufficient amount of nonnative training data would often reduce their efficacy on mispronunciation detection and diagnosis (MD&D). In response, we put forward a novel MD&D method that leverages non-autoregressive (NAR) E2E neural modeling to dramatically speed up the inference time while maintaining performance in line with the conventional E2E neural methods. In addition, we design and develop a pronunciation modeling network stacked on top of the NAR E2E models of our method to further boost the effectiveness of MD&D. Empirical experiments conducted on the L2-ARCTIC English dataset seems to validate the feasibility of our method, in comparison to some top-of-the-line E2E models and an iconic pronunciation-scoring based method built on a DNN-HMM acoustic model. : Accepted for ICASSP2022 Article in Journal/Newspaper Arctic DataCite Metadata Store (German National Library of Science and Technology) Arctic |
institution |
Open Polar |
collection |
DataCite Metadata Store (German National Library of Science and Technology) |
op_collection_id |
ftdatacite |
language |
unknown |
topic |
Computation and Language cs.CL Multimedia cs.MM FOS Computer and information sciences |
spellingShingle |
Computation and Language cs.CL Multimedia cs.MM FOS Computer and information sciences Wang, Hsin-Wei Yan, Bi-Cheng Chiu, Hsuan-Sheng Hsu, Yung-Chang Chen, Berlin Exploring Non-Autoregressive End-To-End Neural Modeling For English Mispronunciation Detection And Diagnosis |
topic_facet |
Computation and Language cs.CL Multimedia cs.MM FOS Computer and information sciences |
description |
End-to-end (E2E) neural modeling has emerged as one predominant school of thought to develop computer-assisted language training (CAPT) systems, showing competitive performance to conventional pronunciation-scoring based methods. However, current E2E neural methods for CAPT are faced with at least two pivotal challenges. On one hand, most of the E2E methods operate in an autoregressive manner with left-to-right beam search to dictate the pronunciations of an L2 learners. This however leads to very slow inference speed, which inevitably hinders their practical use. On the other hand, E2E neural methods are normally data greedy and meanwhile an insufficient amount of nonnative training data would often reduce their efficacy on mispronunciation detection and diagnosis (MD&D). In response, we put forward a novel MD&D method that leverages non-autoregressive (NAR) E2E neural modeling to dramatically speed up the inference time while maintaining performance in line with the conventional E2E neural methods. In addition, we design and develop a pronunciation modeling network stacked on top of the NAR E2E models of our method to further boost the effectiveness of MD&D. Empirical experiments conducted on the L2-ARCTIC English dataset seems to validate the feasibility of our method, in comparison to some top-of-the-line E2E models and an iconic pronunciation-scoring based method built on a DNN-HMM acoustic model. : Accepted for ICASSP2022 |
format |
Article in Journal/Newspaper |
author |
Wang, Hsin-Wei Yan, Bi-Cheng Chiu, Hsuan-Sheng Hsu, Yung-Chang Chen, Berlin |
author_facet |
Wang, Hsin-Wei Yan, Bi-Cheng Chiu, Hsuan-Sheng Hsu, Yung-Chang Chen, Berlin |
author_sort |
Wang, Hsin-Wei |
title |
Exploring Non-Autoregressive End-To-End Neural Modeling For English Mispronunciation Detection And Diagnosis |
title_short |
Exploring Non-Autoregressive End-To-End Neural Modeling For English Mispronunciation Detection And Diagnosis |
title_full |
Exploring Non-Autoregressive End-To-End Neural Modeling For English Mispronunciation Detection And Diagnosis |
title_fullStr |
Exploring Non-Autoregressive End-To-End Neural Modeling For English Mispronunciation Detection And Diagnosis |
title_full_unstemmed |
Exploring Non-Autoregressive End-To-End Neural Modeling For English Mispronunciation Detection And Diagnosis |
title_sort |
exploring non-autoregressive end-to-end neural modeling for english mispronunciation detection and diagnosis |
publisher |
arXiv |
publishDate |
2021 |
url |
https://dx.doi.org/10.48550/arxiv.2111.00844 https://arxiv.org/abs/2111.00844 |
geographic |
Arctic |
geographic_facet |
Arctic |
genre |
Arctic |
genre_facet |
Arctic |
op_rights |
arXiv.org perpetual, non-exclusive license http://arxiv.org/licenses/nonexclusive-distrib/1.0/ |
op_doi |
https://doi.org/10.48550/arxiv.2111.00844 |
_version_ |
1766340685241253888 |