Difference-of-Convex optimization for variational kl-corrected inference in dirichlet process mixtures.
Rasmus BonnevieMikkel N. SchmidtMorten MørupPublished in: MLSP (2017)
Keyphrases
- convex optimization
- dirichlet process
- mixture model
- dirichlet process mixture models
- variational methods
- gaussian processes
- multi task learning
- bayesian model
- variational inference
- generative model
- em algorithm
- expectation maximization
- total variation
- posterior distribution
- image segmentation
- probabilistic model
- bayesian framework
- latent topics
- multi task
- optic flow
- approximate inference
- model selection
- exact inference
- structured prediction
- bayesian inference
- higher order
- probability density function
- image processing
- high order
- maximum likelihood
- unsupervised learning
- missing data
- computer vision