A New Discriminative Kernel from Probabilistic Models.
Koji TsudaMotoaki KawanabeGunnar RätschSören SonnenburgKlaus-Robert MüllerPublished in: Neural Comput. (2002)
Keyphrases
- probabilistic model
- discriminative models
- generative model
- graphical models
- fisher kernel
- conditional random fields
- bayesian inference
- expectation maximization
- language model
- kernel function
- linear classifiers
- hidden variables
- class separability
- mixture model
- crf model
- latent variables
- kernel machines
- semi supervised
- multi class
- kernel regression
- support vector
- discriminative power
- feature extraction
- optimal kernel
- bayesian network classifiers
- discriminative learning
- component analysis
- kernel methods
- topic models
- feature space
- feature selection
- kernel learning
- gaussian processes
- unsupervised learning
- video based face recognition