Norm-Product Belief Propagation: Primal-Dual Message-Passing for Approximate Inference.
Tamir HazanAmnon ShashuaPublished in: IEEE Trans. Inf. Theory (2010)
Keyphrases
- belief propagation
- message passing
- approximate inference
- primal dual
- graphical models
- linear programming
- graph cuts
- linear program
- convex optimization
- markov random field
- factor graphs
- loopy belief propagation
- exact inference
- approximation algorithms
- probabilistic inference
- stereo matching
- markov networks
- free energy
- structured prediction
- pairwise
- fixed point
- energy minimization
- probabilistic graphical models
- bayesian networks
- energy function
- objective function
- image segmentation
- partition function
- statistical physics
- conditional random fields
- markov chain monte carlo
- distributed systems
- probabilistic model
- valid inequalities
- max product
- higher order
- special case