May 2018
Beginner
490 pages
13h 16m
English
An RBM increases in efficiency when random variable sampling is applied during the training process. Gibbs sampling generates a Markov chain of samples that takes the nearby samples into account as well.
Random sampling makes the system more objective. Hidden units remain latent (this is a two-layer system), which confers its dream property.
Read now
Unlock full access