Variational Autoencoder

Training a VAE on MNIST from scratch using Keras.

Projections of Probability Distributions and the Reparameterization Trick

A brief review of M-projection, I-projection, REINFORCE, and the Reparameterization Trick.

Restricted Boltzmann Machine and Contrastive Divergence

RBM training with CD from scratch on MNIST data.

Gibbs Sampling in Pairwise Markov Networks

Discrete pairwise Markov networks are pretty straightforward to work with analytically. Let's use the canonical overcomplete representation, where the sufficient statistics functions are indicator functions …