Speech Recognition for Task Domains with Sparse Matched Training Data
We propose two approaches to handle speech recognition for task domains with sparse matched training data. One is an active learning method that selects training data for the target domain from another general domain that already has a significant amount of labeled speech data. This method uses attr...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2020-09-01
|
Series: | Applied Sciences |
Subjects: | |
Online Access: | https://www.mdpi.com/2076-3417/10/18/6155 |
_version_ | 1797554605835419648 |
---|---|
author | Byung Ok Kang Hyeong Bae Jeon Jeon Gue Park |
author_facet | Byung Ok Kang Hyeong Bae Jeon Jeon Gue Park |
author_sort | Byung Ok Kang |
collection | DOAJ |
description | We propose two approaches to handle speech recognition for task domains with sparse matched training data. One is an active learning method that selects training data for the target domain from another general domain that already has a significant amount of labeled speech data. This method uses attribute-disentangled latent variables. For the active learning process, we designed an integrated system consisting of a variational autoencoder with an encoder that infers latent variables with disentangled attributes from the input speech, and a classifier that selects training data with attributes matching the target domain. The other method combines data augmentation methods for generating matched target domain speech data and transfer learning methods based on teacher/student learning. To evaluate the proposed method, we experimented with various task domains with sparse matched training data. The experimental results show that the proposed method has qualitative characteristics that are suitable for the desired purpose, it outperforms random selection, and is comparable to using an equal amount of additional target domain data. |
first_indexed | 2024-03-10T16:34:28Z |
format | Article |
id | doaj.art-fcc47a5bbbac4264af1f90632a62b6cb |
institution | Directory Open Access Journal |
issn | 2076-3417 |
language | English |
last_indexed | 2024-03-10T16:34:28Z |
publishDate | 2020-09-01 |
publisher | MDPI AG |
record_format | Article |
series | Applied Sciences |
spelling | doaj.art-fcc47a5bbbac4264af1f90632a62b6cb2023-11-20T12:37:25ZengMDPI AGApplied Sciences2076-34172020-09-011018615510.3390/app10186155Speech Recognition for Task Domains with Sparse Matched Training DataByung Ok Kang0Hyeong Bae Jeon1Jeon Gue Park2Electronics and Telecommunications Research Institute, Daejeon 34129, KoreaElectronics and Telecommunications Research Institute, Daejeon 34129, KoreaElectronics and Telecommunications Research Institute, Daejeon 34129, KoreaWe propose two approaches to handle speech recognition for task domains with sparse matched training data. One is an active learning method that selects training data for the target domain from another general domain that already has a significant amount of labeled speech data. This method uses attribute-disentangled latent variables. For the active learning process, we designed an integrated system consisting of a variational autoencoder with an encoder that infers latent variables with disentangled attributes from the input speech, and a classifier that selects training data with attributes matching the target domain. The other method combines data augmentation methods for generating matched target domain speech data and transfer learning methods based on teacher/student learning. To evaluate the proposed method, we experimented with various task domains with sparse matched training data. The experimental results show that the proposed method has qualitative characteristics that are suitable for the desired purpose, it outperforms random selection, and is comparable to using an equal amount of additional target domain data.https://www.mdpi.com/2076-3417/10/18/6155automatic speech recognitionsparse training datadeep neural networkactive learningtransfer learning |
spellingShingle | Byung Ok Kang Hyeong Bae Jeon Jeon Gue Park Speech Recognition for Task Domains with Sparse Matched Training Data Applied Sciences automatic speech recognition sparse training data deep neural network active learning transfer learning |
title | Speech Recognition for Task Domains with Sparse Matched Training Data |
title_full | Speech Recognition for Task Domains with Sparse Matched Training Data |
title_fullStr | Speech Recognition for Task Domains with Sparse Matched Training Data |
title_full_unstemmed | Speech Recognition for Task Domains with Sparse Matched Training Data |
title_short | Speech Recognition for Task Domains with Sparse Matched Training Data |
title_sort | speech recognition for task domains with sparse matched training data |
topic | automatic speech recognition sparse training data deep neural network active learning transfer learning |
url | https://www.mdpi.com/2076-3417/10/18/6155 |
work_keys_str_mv | AT byungokkang speechrecognitionfortaskdomainswithsparsematchedtrainingdata AT hyeongbaejeon speechrecognitionfortaskdomainswithsparsematchedtrainingdata AT jeonguepark speechrecognitionfortaskdomainswithsparsematchedtrainingdata |