Topic tracking language model for speech recognition

Shinji Watanabe, Tomoharu Iwata, Takaaki Hori, Atsushi Sako, Yasuo Ariki

Research output: Contribution to journalArticle

25 Citations (Scopus)

Abstract

In a real environment, acoustic and language features often vary depending on the speakers, speaking styles and topic changes. To accommodate these changes, speech recognition approaches that include the incremental tracking of changing environments have attracted attention. This paper proposes a topic tracking language model that can adaptively track changes in topics based on current text information and previously estimated topic models in an on-line manner. The proposed model is applied to language model adaptation in speech recognition. We use the MIT OpenCourseWare corpus and Corpus of Spontaneous Japanese in speech recognition experiments, and show the effectiveness of the proposed method.

Original languageEnglish
Pages (from-to)440-461
Number of pages22
JournalComputer Speech and Language
Volume25
Issue number2
DOIs
Publication statusPublished - 2011 Apr
Externally publishedYes

Fingerprint

Language Model
Speech Recognition
Speech recognition
Acoustics
Vary
Model
Experiment
Corpus
Experiments

Keywords

  • Language model
  • Latent topic model
  • On-line algorithm
  • Speech recognition
  • Topic tracking

ASJC Scopus subject areas

  • Software
  • Human-Computer Interaction
  • Theoretical Computer Science

Cite this

Topic tracking language model for speech recognition. / Watanabe, Shinji; Iwata, Tomoharu; Hori, Takaaki; Sako, Atsushi; Ariki, Yasuo.

In: Computer Speech and Language, Vol. 25, No. 2, 04.2011, p. 440-461.

Research output: Contribution to journalArticle

Watanabe, Shinji ; Iwata, Tomoharu ; Hori, Takaaki ; Sako, Atsushi ; Ariki, Yasuo. / Topic tracking language model for speech recognition. In: Computer Speech and Language. 2011 ; Vol. 25, No. 2. pp. 440-461.
@article{f044200caf554467ab7f59ee00949395,
title = "Topic tracking language model for speech recognition",
abstract = "In a real environment, acoustic and language features often vary depending on the speakers, speaking styles and topic changes. To accommodate these changes, speech recognition approaches that include the incremental tracking of changing environments have attracted attention. This paper proposes a topic tracking language model that can adaptively track changes in topics based on current text information and previously estimated topic models in an on-line manner. The proposed model is applied to language model adaptation in speech recognition. We use the MIT OpenCourseWare corpus and Corpus of Spontaneous Japanese in speech recognition experiments, and show the effectiveness of the proposed method.",
keywords = "Language model, Latent topic model, On-line algorithm, Speech recognition, Topic tracking",
author = "Shinji Watanabe and Tomoharu Iwata and Takaaki Hori and Atsushi Sako and Yasuo Ariki",
year = "2011",
month = "4",
doi = "10.1016/j.csl.2010.07.006",
language = "English",
volume = "25",
pages = "440--461",
journal = "Computer Speech and Language",
issn = "0885-2308",
publisher = "Academic Press Inc.",
number = "2",

}

TY - JOUR

T1 - Topic tracking language model for speech recognition

AU - Watanabe, Shinji

AU - Iwata, Tomoharu

AU - Hori, Takaaki

AU - Sako, Atsushi

AU - Ariki, Yasuo

PY - 2011/4

Y1 - 2011/4

N2 - In a real environment, acoustic and language features often vary depending on the speakers, speaking styles and topic changes. To accommodate these changes, speech recognition approaches that include the incremental tracking of changing environments have attracted attention. This paper proposes a topic tracking language model that can adaptively track changes in topics based on current text information and previously estimated topic models in an on-line manner. The proposed model is applied to language model adaptation in speech recognition. We use the MIT OpenCourseWare corpus and Corpus of Spontaneous Japanese in speech recognition experiments, and show the effectiveness of the proposed method.

AB - In a real environment, acoustic and language features often vary depending on the speakers, speaking styles and topic changes. To accommodate these changes, speech recognition approaches that include the incremental tracking of changing environments have attracted attention. This paper proposes a topic tracking language model that can adaptively track changes in topics based on current text information and previously estimated topic models in an on-line manner. The proposed model is applied to language model adaptation in speech recognition. We use the MIT OpenCourseWare corpus and Corpus of Spontaneous Japanese in speech recognition experiments, and show the effectiveness of the proposed method.

KW - Language model

KW - Latent topic model

KW - On-line algorithm

KW - Speech recognition

KW - Topic tracking

UR - http://www.scopus.com/inward/record.url?scp=78049521179&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78049521179&partnerID=8YFLogxK

U2 - 10.1016/j.csl.2010.07.006

DO - 10.1016/j.csl.2010.07.006

M3 - Article

AN - SCOPUS:78049521179

VL - 25

SP - 440

EP - 461

JO - Computer Speech and Language

JF - Computer Speech and Language

SN - 0885-2308

IS - 2

ER -