Parallelizable Sampling of Markov Random Fields.
James MartensIlya SutskeverPublished in: AISTATS (2010)
Keyphrases
- markov random field
- belief propagation
- pairwise
- parameter estimation
- mrf model
- random fields
- higher order
- maximum a posteriori
- energy function
- graph cuts
- energy minimization
- image segmentation
- low level vision
- image restoration
- mrf models
- potential functions
- conditional random fields
- potts model
- partition function
- efficient inference
- loopy belief propagation
- random sampling
- message passing
- iterative conditional
- piecewise constant functions
- textured images
- markov chain monte carlo
- approximate inference
- monte carlo
- map inference
- sample size
- label field