October 2018
Beginner
362 pages
9h 32m
English
Fully visible belief networks are a class of explicit density models and a form of deep belief network. They use the chain rule to decompose a probability distribution
over a vector, into a product over each of the members of the vector, represented between by
. All together, it's formula is:

The most popular model in this family is PixelCNN, an autoregressive generative model. Pixels approach image generation problems ...
Read now
Unlock full access