Probabilistic Linear Discriminant Analysis for Acoustic Modeling

Liang Lu, Steve Renals

Research output: Contribution to journalArticlepeer-review


In this letter, we propose a new acoustic modeling approach for automatic speech recognition based on probabilistic linear discriminant analysis (PLDA), which is used to model the state density function for the standard hidden Markov models (HMMs). Unlike the conventional Gaussian mixture models (GMMs) where the correlations are weakly modelled by using the diagonal covariance matrices, PLDA captures the correlations of feature vector in subspaces without vastly expanding the model. It also allows the usage of high dimensional feature input, and therefore is more flexible to make use of different type of acoustic features. We performed the preliminary experiments on the Switchboard corpus, and demonstrated the feasibility of this acoustic model.
Original languageEnglish
Pages (from-to)702-706
Number of pages5
JournalIEEE Signal Processing Letters
Issue number6
Publication statusPublished - 1 Jun 2014


  • covariance matrices
  • hidden Markov models
  • probability
  • speech recognition
  • GMM
  • HMM
  • PLDA
  • acoustic modeling approach
  • automatic speech recognition
  • conventional Gaussian mixture models
  • diagonal covariance matrices
  • feature vector
  • probabilistic linear discriminant analysis
  • speech recognition systems
  • standard hidden Markov models
  • state density function
  • Analytical models
  • Computational modeling
  • Hidden Markov models
  • Mel frequency cepstral coefficient
  • Speech recognition
  • Training
  • Acoustic modeling


Dive into the research topics of 'Probabilistic Linear Discriminant Analysis for Acoustic Modeling'. Together they form a unique fingerprint.

Cite this