A recommender system based on Belief Propagation over Pairwise Markov Random Fields.
Erman AydayJun ZouArash EinolghozatiFaramarz FekriPublished in: Allerton Conference (2012)
Keyphrases
- belief propagation
- markov random field
- pairwise
- message passing
- graph cuts
- approximate inference
- higher order
- random fields
- mrf model
- fixed point
- maximum a posteriori
- energy function
- parameter estimation
- loopy belief propagation
- energy minimization
- image restoration
- potential functions
- factor graphs
- image segmentation
- conditional random fields
- markov networks
- similarity measure
- partition function
- semi supervised
- potts model
- low level vision
- statistical physics
- graphical models
- probabilistic model
- three dimensional