Principled Parallel Mean-Field Inference for Discrete Random Fields.
Pierre BaquéTimur M. BagautdinovFrançois FleuretPascal FuaPublished in: CVPR (2016)
Keyphrases
- random fields
- markov random field
- partition function
- markov processes
- gibbs sampler
- conditional random fields
- maximum entropy
- random field models
- loopy belief propagation
- parameter estimation
- non stationary
- pseudo likelihood
- random field model
- higher order
- belief propagation
- bayesian inference
- graph cuts
- probabilistic model
- pairwise
- energy minimization
- maximum a posteriori
- closed form
- energy function
- nonparametric density estimation
- feature selection
- belief networks
- exact inference
- markov networks
- prior information
- bayesian networks