Maximum conditional likelihood linear regression and maximum a posteriori for hidden conditional random fields speaker adaptation.
Yun-Hsuan SungConstantinos BoulisDaniel JurafskyPublished in: ICASSP (2008)
Keyphrases
- linear regression
- maximum a posteriori
- maximum likelihood
- markov random field
- least squares
- image reconstruction
- markov networks
- bayesian framework
- em algorithm
- hyperparameters
- expectation maximization
- posterior distribution
- energy function
- prior distribution
- speech recognition
- logistic regression
- posterior probability
- belief propagation
- non stationary
- super resolution
- higher order
- pattern recognition