Empirical Analysis of the Divergence of Gibbs Sampling Based Learning Algorithms for Restricted Boltzmann Machines.
Asja FischerChristian IgelPublished in: ICANN (3) (2010)
Keyphrases
- empirical analysis
- restricted boltzmann machine
- learning algorithm
- monte carlo
- deep learning
- markov chain monte carlo
- contrastive divergence
- theoretical analysis
- empirical studies
- conditional random fields
- probabilistic graphical models
- markov random field
- loopy belief propagation
- deep belief networks
- semi supervised
- graphical models
- optical flow
- training data
- posterior distribution
- machine learning
- higher order
- learning rate
- hidden layer
- reinforcement learning
- latent variables
- transfer learning
- machine learning algorithms
- probabilistic model
- learning process