Adaptation approaches for pronunciation scoring with sparse training data
In Computer Assisted Language Learning systems, pronunciation scoring consists in providing a score grading the overall pronunciation quality of the speech uttered by a student. In this work, a log-likelihood ratio obtained with respect to two automatic speech recognition (ASR) models was used as sc...
Guardado en:
Publicado: |
2017
|
---|---|
Materias: | |
Acceso en línea: | https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_03029743_v10458LNAI_n_p87_Landini http://hdl.handle.net/20.500.12110/paper_03029743_v10458LNAI_n_p87_Landini |
Aporte de: |
id |
paper:paper_03029743_v10458LNAI_n_p87_Landini |
---|---|
record_format |
dspace |
spelling |
paper:paper_03029743_v10458LNAI_n_p87_Landini2023-06-08T15:28:15Z Adaptation approaches for pronunciation scoring with sparse training data Computer-assisted language learning Log-likelihood ratio MAP adaptation Pronunciation scoring Computer aided instruction E-learning Grading Linguistics Automatic speech recognition Computer assisted language learning Computer assisted language learning systems Log likelihood ratio MAP adaptation Pronunciation quality Pronunciation scoring Scoring performance Speech recognition In Computer Assisted Language Learning systems, pronunciation scoring consists in providing a score grading the overall pronunciation quality of the speech uttered by a student. In this work, a log-likelihood ratio obtained with respect to two automatic speech recognition (ASR) models was used as score. One model represents native pronunciation while the other one captures non-native pronunciation. Different approaches to obtain each model and different amounts of training data were analyzed. The best results were obtained training an ASR system using a separate large corpus without pronunciation quality annotations and then adapting it to the native and non-native data, sequentially. Nevertheless, when models are trained directly on the native and non-native data, pronunciation scoring performance is similar. This is a surprising result considering that word error rates for these models are significantly worse, indicating that ASR performance is not a good predictor of pronunciation scoring performance on this system. © Springer International Publishing AG 2017. 2017 https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_03029743_v10458LNAI_n_p87_Landini http://hdl.handle.net/20.500.12110/paper_03029743_v10458LNAI_n_p87_Landini |
institution |
Universidad de Buenos Aires |
institution_str |
I-28 |
repository_str |
R-134 |
collection |
Biblioteca Digital - Facultad de Ciencias Exactas y Naturales (UBA) |
topic |
Computer-assisted language learning Log-likelihood ratio MAP adaptation Pronunciation scoring Computer aided instruction E-learning Grading Linguistics Automatic speech recognition Computer assisted language learning Computer assisted language learning systems Log likelihood ratio MAP adaptation Pronunciation quality Pronunciation scoring Scoring performance Speech recognition |
spellingShingle |
Computer-assisted language learning Log-likelihood ratio MAP adaptation Pronunciation scoring Computer aided instruction E-learning Grading Linguistics Automatic speech recognition Computer assisted language learning Computer assisted language learning systems Log likelihood ratio MAP adaptation Pronunciation quality Pronunciation scoring Scoring performance Speech recognition Adaptation approaches for pronunciation scoring with sparse training data |
topic_facet |
Computer-assisted language learning Log-likelihood ratio MAP adaptation Pronunciation scoring Computer aided instruction E-learning Grading Linguistics Automatic speech recognition Computer assisted language learning Computer assisted language learning systems Log likelihood ratio MAP adaptation Pronunciation quality Pronunciation scoring Scoring performance Speech recognition |
description |
In Computer Assisted Language Learning systems, pronunciation scoring consists in providing a score grading the overall pronunciation quality of the speech uttered by a student. In this work, a log-likelihood ratio obtained with respect to two automatic speech recognition (ASR) models was used as score. One model represents native pronunciation while the other one captures non-native pronunciation. Different approaches to obtain each model and different amounts of training data were analyzed. The best results were obtained training an ASR system using a separate large corpus without pronunciation quality annotations and then adapting it to the native and non-native data, sequentially. Nevertheless, when models are trained directly on the native and non-native data, pronunciation scoring performance is similar. This is a surprising result considering that word error rates for these models are significantly worse, indicating that ASR performance is not a good predictor of pronunciation scoring performance on this system. © Springer International Publishing AG 2017. |
title |
Adaptation approaches for pronunciation scoring with sparse training data |
title_short |
Adaptation approaches for pronunciation scoring with sparse training data |
title_full |
Adaptation approaches for pronunciation scoring with sparse training data |
title_fullStr |
Adaptation approaches for pronunciation scoring with sparse training data |
title_full_unstemmed |
Adaptation approaches for pronunciation scoring with sparse training data |
title_sort |
adaptation approaches for pronunciation scoring with sparse training data |
publishDate |
2017 |
url |
https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_03029743_v10458LNAI_n_p87_Landini http://hdl.handle.net/20.500.12110/paper_03029743_v10458LNAI_n_p87_Landini |
_version_ |
1768541844910637056 |