A segment based probabilistic generative model of speech.

Kannan Achan Sam T. Roweis Aaron Hertzmann Brendan J. Frey

Published in: ICASSP (5) (2005)

Keyphrases

generative model
probabilistic model
posterior probability
prior knowledge
mixture model
bayesian framework

maximum entropy principle
object categories
speech recognition
topic models
expectation maximization
semi supervised

em algorithm
latent dirichlet allocation
discriminative learning
learned models
discriminative models
conditional probabilities

training set
video sequences
bayesian inference
latent space
video representation
data sets