Feature space maximum a posteriori linear regression for adaptation of deep neural networks.
Zhen HuangJinyu LiSabato Marco SiniscalchiI-Fan ChenChao WengChin-Hui LeePublished in: INTERSPEECH (2014)
Keyphrases
- linear regression
- maximum a posteriori
- feature space
- map estimation
- markov random field
- least squares
- maximum likelihood
- image reconstruction
- bayesian framework
- hyperparameters
- regression methods
- posterior distribution
- prior model
- energy function
- expectation maximization
- data points
- feature vectors
- support vector machine
- high dimensional
- em algorithm
- nonlinear regression
- label field
- kernel function
- feature selection
- low dimensional
- prior distribution
- training set
- classification accuracy
- principal component analysis
- input space
- feature extraction
- kernel regression
- dimensionality reduction
- generalized gaussian
- computer vision
- machine learning
- edge detection
- graphical models
- regression method
- gaussian mixture model