Phase transitions in Restricted Boltzmann Machines with generic priors.
Adriano BarraGiuseppe GenovesePeter SollichDaniele TantariPublished in: CoRR (2016)
Keyphrases
- phase transition
- restricted boltzmann machine
- constraint satisfaction
- satisfiability problem
- deep learning
- random constraint satisfaction problems
- contrastive divergence
- learning algorithm
- randomly generated
- np complete
- random instances
- cellular automata
- conditional random fields
- deep belief networks
- probabilistic graphical models
- hidden layer
- machine learning
- loopy belief propagation
- prior knowledge
- semi supervised
- information extraction
- bayesian framework
- graph cuts
- constraint satisfaction problems