The Resource Ensemble acoustic modeling in automatic speech recognition

Ensemble acoustic modeling in automatic speech recognition

Label
Ensemble acoustic modeling in automatic speech recognition
Title
Ensemble acoustic modeling in automatic speech recognition
Creator
Author
Subject
Language
eng
Summary
In this dissertation, several new approaches of using data sampling to construct an Ensemble of Acoustic Models (EAM) for speech recognition are proposed. A straightforward method of data sampling is Cross Validation (CV) data partition. In the direction of improving inter-model diversity within an EAM for speaker independent speech recognition, we propose Speaker Clustering (SC) based data sampling. In the direction of improving base model quality as well as inter-model diversity, we further investigate the effects of several successful techniques of single model training in speech recognition on the proposed ensemble acoustic models, including Cross Validation Expectation Maximization (CVEM), Discriminative Training (DT), and Multiple Layer Perceptron (MLP) features. We have evaluated the proposed methods on TIMIT phoneme recognition task as well as on a telemedicine automatic captioning task. The proposed EAMs have led to significant improvements in recognition accuracy over conventional Hidden Markov Model (HMM) baseline systems, and the integration of EAM with CVEM, DT and MLP has also significantly improved the accuracy performances of CVEM, DT, and MLP based single model systems. We further investigated the largely unstudied factor of inter-model diversity, and proposed several methods to explicit measure inter-model diversity. We demonstrate a positive relation between enlarging inter-model diversity and increasing EAM quality. Compacting the acoustic model to a reasonable size for practical applications while maintaining a reasonable performance is needed for EAM. Toward this goal, in this dissertation, we discuss and investigate several distance measures and proposed global optimization algorithms for clustering methods. We also proposed an explicit PDT (EPDT) state tying approach that allows Phoneme data Sharing (PS) for its potential capability in accommodating pronunciation variations
Cataloging source
MUU
http://library.link/vocab/creatorDate
1983-
http://library.link/vocab/creatorName
Chen, Xin
Degree
Ph. D.
Dissertation note
Thesis
Dissertation year
2011.
Government publication
government publication of a state province territory dependency etc
Granting institution
University of Missouri--Columbia,
Index
no index present
Literary form
non fiction
Nature of contents
dictionaries
Label
Ensemble acoustic modeling in automatic speech recognition
Instantiates
Publication
Contributor
Thesis advisor
Note
Advisor: Yunxin Zhao
Carrier category
online resource
Carrier category code
  • cr
Carrier MARC source
rdacarrier.
Content category
text
Content type code
  • txt
Content type MARC source
rdacontent.
Control code
872561309
Extent
1 online resource (xiii, 106 pages)
Form of item
online
Media category
computer
Media MARC source
rdamedia.
Media type code
  • c
Specific material designation
remote
System control number
(OCoLC)872561309
Label
Ensemble acoustic modeling in automatic speech recognition
Publication
Contributor
Thesis advisor
Note
Advisor: Yunxin Zhao
Carrier category
online resource
Carrier category code
  • cr
Carrier MARC source
rdacarrier.
Content category
text
Content type code
  • txt
Content type MARC source
rdacontent.
Control code
872561309
Extent
1 online resource (xiii, 106 pages)
Form of item
online
Media category
computer
Media MARC source
rdamedia.
Media type code
  • c
Specific material designation
remote
System control number
(OCoLC)872561309

Library Locations

  • Health Sciences LibraryBorrow it
    2411 Holmes St, Kansas City, Kansas City, MO, 64108, US
    39.083418 -94.575323
  • LaBudde Special CollectionsBorrow it
    800 E 51st St, Kansas City, MO, 64110, US
    39.034642 -94.576835
  • Leon E. Bloch Law LibraryBorrow it
    500 E. 52nd Street, Kansas City, MO, 64110, US
    39.032488 -94.581967
  • Marr Sound ArchivesBorrow it
    800 E 51st St, Kansas City, MO, 64110, US
    39.034642 -94.576835
  • Miller Nichols LibraryBorrow it
    800 E 51st St, Kansas City, MO, 64110, US
    39.035061 -94.576518
  • UMKCBorrow it
    800 E 51st St, Kansas City, MO, 64110, US
    39.035061 -94.576518
  • UMKCBorrow it
    800 E 51st St, Kansas City, MO, 64110, US
Processing Feedback ...