Generating small, accurate acoustic models with a modified bayesian information criterion

Kai Yu, Robin A Rutenbar

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

Abstract

Although Gaussian mixture models are commonly used in acoustic models for speech recognition, there is no standard method for determining the number of mixture components. Most models arbitrarily assign the number of mixture components with little justification. While model selection techniques with a mathematical derivation, such as the Bayesian information criterion (BIC), have been applied, these criteria focus on properly modeling the true distribution of individual tied-states (senones) without considering the entire acoustic model; this leads to suboptimal speech recognition performance. In this paper we present a method to generate statistically-justified acoustic models that consider inter-senone effects by modifying the BIC. Experimental results in the CMU Communicator domain show that in contrast to previous strategies, the new method generates not only attractively smaller acoustic models, but also ones with lower word error rate.

Original languageEnglish (US)
Title of host publicationInternational Speech Communication Association - 8th Annual Conference of the International Speech Communication Association, Interspeech 2007
Pages1165-1168
Number of pages4
StatePublished - Dec 1 2007
Event8th Annual Conference of the International Speech Communication Association, Interspeech 2007 - Antwerp, Belgium
Duration: Aug 27 2007Aug 31 2007

Publication series

NameProceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
Volume2
ISSN (Electronic)1990-9772

Other

Other8th Annual Conference of the International Speech Communication Association, Interspeech 2007
CountryBelgium
CityAntwerp
Period8/27/078/31/07

Keywords

  • Acoustic model training
  • BIC
  • Gaussian mixture models
  • Model selection

ASJC Scopus subject areas

  • Computer Science Applications
  • Software
  • Modeling and Simulation
  • Linguistics and Language
  • Communication

Fingerprint Dive into the research topics of 'Generating small, accurate acoustic models with a modified bayesian information criterion'. Together they form a unique fingerprint.

Cite this