Deep learning via message passing algorithms based on belief propagation.
Carlo LucibelloFabrizio PittorinoGabriele PeruginiRiccardo ZecchinaPublished in: CoRR (2021)
Keyphrases
- message passing
- belief propagation
- graphical models
- max product
- loopy belief propagation
- inference in graphical models
- clique potentials
- deep learning
- markov random field
- approximate inference
- stereo matching
- graph cuts
- probabilistic inference
- factor graphs
- pairwise
- sum product
- distributed systems
- energy minimization
- energy function
- markov networks
- generalized belief propagation
- tree reweighted
- free energy
- preprocessing
- junction tree
- probabilistic model
- multiscale
- bayesian networks
- probabilistic graphical models
- image registration