Boltzmann Machines are bidirectionally connected networks of stochastic processing units. It can be used to learn important aspects of an unknown probability distribution based on samples from the distribution. We explore Gibbs Sampling and Contrastive Divergence.
