Incremental MAP estimation of HMMs for efficient training and improved performance

  title={Incremental MAP estimation of HMMs for efficient training and improved performance},
  author={Yoshihoko Goto and Mike Hochberg and Daniel J. Mashao and Harvey F. Silverman},
Continuous density observation hidden Markov models (CD-HMMs) have been shown to perform better than their discrete counterparts. However, because the observation distribution is usually represented with a mixture of multi-variate normal densities, the training time for a CD-HMM can be prohibitively long. This paper presents a new approach to speed-up the convergence of CD-HMM training using a stochastic, incremental variant of the EM algorithm. The algorithm randomly selects a subset of data… CONTINUE READING


Publications citing this paper.


Publications referenced by this paper.
Showing 1-9 of 9 references

Connectionist Speech Recognition: A Hybrid Approach

  • Herv e A. Bourlard, Nelson Morgan
  • Kluwer Aca- demic Publishers,
  • 1994
1 Excerpt

On the implementation and computation of training an HMM recognizer having explicit state durations and multiple- feature-set tied-mixture observation probabilities

  • Harvey F. Silverman, Yoshihoko Gotoh
  • Tech- nical Report LEMS Monograph Series: 1-1,
  • 1994
3 Excerpts

Silverman and Yoshihoko Gotoh . On theimplementation and computation of training an HMMrecognizer having explicit state durations and multiplefeatureset tiedmixture observation probabilities

  • F. Harvey
  • 1994

A new view ofthe EM algorithm that justi es incremental and othervariants

  • Radford M. Neal, Geo rey E. Hinton.

Similar Papers

Loading similar papers…