Downloads: 95

Files in This Item:
File Description SizeFormat 
transinf.2015EDP7047.pdf962.31 kBAdobe PDFView/Open
Title: Automatic lecture transcription based on discriminative data selection for lightly supervised acoustic model training
Authors: Li, Sheng
Akita, Yuya  kyouindb  KAKEN_id
Kawahara, Tatsuya  kyouindb  KAKEN_id  orcid https://orcid.org/0000-0002-2686-2296 (unconfirmed)
Author's alias: 河原, 達也
秋田, 祐哉
Keywords: speech recognition
acoustic model
lightly supervised training
lecture transcription
Issue Date: 2015
Publisher: Institute of Electronics, Information and Communication Engineers(IEICE)
Journal title: IEICE Transactions on Information and Systems
Volume: E98.D
Issue: 8
Start page: 1545
End page: 1552
Abstract: The paper addresses a scheme of lightly supervised training of an acoustic model, which exploits a large amount of data with closed caption texts but not faithful transcripts. In the proposed scheme, a sequence of the closed caption text and that of the ASR hypothesis by the baseline system are aligned. Then, a set of dedicated classifiers is designed and trained to select the correct one among them or reject both. It is demonstrated that the classifiers can effectively filter the usable data for acoustic model training. The scheme realizes automatic training of the acoustic model with an increased amount of data. A significant improvement in the ASR accuracy is achieved from the baseline system and also in comparison with the conventional method of lightly supervised training based on simple matching.
Rights: Copyright © 2015 The Institute of Electronics, Information and Communication Engineers.
URI: http://hdl.handle.net/2433/219407
DOI(Published Version): 10.1587/transinf.2015EDP7047
Appears in Collections:Journal Articles

Show full item record

Export to RefWorks


Export Format: 


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.