Hidden Markov model and neural network hybrid

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

When there is a mismatch between training and testing environments, statistical pattern classification methods may suffer from severe degradation in their performance because the parameters in the classifiers do not represent the testing data well. The mismatch is typically due to the interference or noises from operating environments. In this paper, a neural network based transformation approach is studied to handle the distribution mismatches between training and testing data. The probability density functions of the statistical classifiers are used as the objective function of the neural network. The neural network maximizes the likelihood of the data from a testing environment, and allows global optimization of the network when used with the statistical pattern classifiers. The proposed approach is applied to the area of automatic speech recognition to recognize noisy distant-talking speech and it reduces the error rate by 52.9%.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages196-203
Number of pages8
Volume2510 LNCS
ISBN (Print)3540000283, 9783540000280
Publication statusPublished - 2002 Jan 1
Event1st EurAsian Conference on Advances in Information and Communication Technology, EurAsia-ICT 2002 - Shiraz, Iran, Islamic Republic of
Duration: 2002 Oct 292002 Oct 31

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume2510 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other1st EurAsian Conference on Advances in Information and Communication Technology, EurAsia-ICT 2002
CountryIran, Islamic Republic of
CityShiraz
Period02/10/2902/10/31

Fingerprint

Hidden Markov models
Markov Model
Neural Networks
Neural networks
Testing
Classifiers
Classifier
Automatic Speech Recognition
Pattern Classification
Global optimization
Speech recognition
Global Optimization
Probability density function
Pattern recognition
Error Rate
Likelihood
Degradation
Objective function
Interference
Maximise

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Yook, D. (2002). Hidden Markov model and neural network hybrid. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 2510 LNCS, pp. 196-203). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2510 LNCS). Springer Verlag.

Hidden Markov model and neural network hybrid. / Yook, Dongsuk.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 2510 LNCS Springer Verlag, 2002. p. 196-203 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2510 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Yook, D 2002, Hidden Markov model and neural network hybrid. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 2510 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 2510 LNCS, Springer Verlag, pp. 196-203, 1st EurAsian Conference on Advances in Information and Communication Technology, EurAsia-ICT 2002, Shiraz, Iran, Islamic Republic of, 02/10/29.
Yook D. Hidden Markov model and neural network hybrid. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 2510 LNCS. Springer Verlag. 2002. p. 196-203. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Yook, Dongsuk. / Hidden Markov model and neural network hybrid. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 2510 LNCS Springer Verlag, 2002. pp. 196-203 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{2bc1574482a04769839d6f0c2241d6ed,
title = "Hidden Markov model and neural network hybrid",
abstract = "When there is a mismatch between training and testing environments, statistical pattern classification methods may suffer from severe degradation in their performance because the parameters in the classifiers do not represent the testing data well. The mismatch is typically due to the interference or noises from operating environments. In this paper, a neural network based transformation approach is studied to handle the distribution mismatches between training and testing data. The probability density functions of the statistical classifiers are used as the objective function of the neural network. The neural network maximizes the likelihood of the data from a testing environment, and allows global optimization of the network when used with the statistical pattern classifiers. The proposed approach is applied to the area of automatic speech recognition to recognize noisy distant-talking speech and it reduces the error rate by 52.9{\%}.",
author = "Dongsuk Yook",
year = "2002",
month = "1",
day = "1",
language = "English",
isbn = "3540000283",
volume = "2510 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "196--203",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Hidden Markov model and neural network hybrid

AU - Yook, Dongsuk

PY - 2002/1/1

Y1 - 2002/1/1

N2 - When there is a mismatch between training and testing environments, statistical pattern classification methods may suffer from severe degradation in their performance because the parameters in the classifiers do not represent the testing data well. The mismatch is typically due to the interference or noises from operating environments. In this paper, a neural network based transformation approach is studied to handle the distribution mismatches between training and testing data. The probability density functions of the statistical classifiers are used as the objective function of the neural network. The neural network maximizes the likelihood of the data from a testing environment, and allows global optimization of the network when used with the statistical pattern classifiers. The proposed approach is applied to the area of automatic speech recognition to recognize noisy distant-talking speech and it reduces the error rate by 52.9%.

AB - When there is a mismatch between training and testing environments, statistical pattern classification methods may suffer from severe degradation in their performance because the parameters in the classifiers do not represent the testing data well. The mismatch is typically due to the interference or noises from operating environments. In this paper, a neural network based transformation approach is studied to handle the distribution mismatches between training and testing data. The probability density functions of the statistical classifiers are used as the objective function of the neural network. The neural network maximizes the likelihood of the data from a testing environment, and allows global optimization of the network when used with the statistical pattern classifiers. The proposed approach is applied to the area of automatic speech recognition to recognize noisy distant-talking speech and it reduces the error rate by 52.9%.

UR - http://www.scopus.com/inward/record.url?scp=84896928614&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84896928614&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84896928614

SN - 3540000283

SN - 9783540000280

VL - 2510 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 196

EP - 203

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

PB - Springer Verlag

ER -