Bayesian large margin hidden Markov models for speech recognition

Jung Chun Chen*, Jen-Tzung Chien

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

8 Scopus citations

Abstract

This paper presents a Bayesian learning approach to large margin classifier for hidden Markov model (HMM) based speech recognition. We build the Bayesian large margin HMMs (BLM-HMMs) and improve the model generalization for handling unknown test environments. Using BLM-HMMs, the variational Bayesian HMM parameters are estimated by maximizing lower bound of a marginal likelihood over the uncertainties of HMM parameters. The Bayesian large margin estimation is performed with frame selection mechanism, and is illustrated to meet the objective of support vector machines, i.e. maximal class margin and minimal training errors. The new objective function is not only interpreted as a discriminative criterion, but also feasible to deal with model selection and adaptive training. Experiments on phone recognition show that BLM-HMMs perform better than other generative and discriminative models.

Original languageEnglish
Title of host publication2009 IEEE International Conference on Acoustics, Speech, and Signal Processing - Proceedings, ICASSP 2009
Pages3765-3768
Number of pages4
DOIs
StatePublished - 23 Sep 2009
Event2009 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2009 - Taipei, Taiwan
Duration: 19 Apr 200924 Apr 2009

Publication series

NameICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
ISSN (Print)1520-6149

Conference

Conference2009 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2009
CountryTaiwan
CityTaipei
Period19/04/0924/04/09

Keywords

  • Bayesian learning
  • Hidden Markov models
  • Large margin classifier
  • Model generalization

Fingerprint Dive into the research topics of 'Bayesian large margin hidden Markov models for speech recognition'. Together they form a unique fingerprint.

Cite this