On the Anatomy of MCMC-Based Maximum Likelihood Learning of Energy-Based Models

Understand

This study investigates the effects of Markov chain Monte Carlo (MCMC) sampling in unsupervised Maximum Likelihood (ML) learning.

• Our attention is restricted to the family of unnormalized probability densities for which the negative log density (or energy function) is a ConvNet.
• We find that many of the techniques used to stabilize training in previous studies are not necessary.
• ML learning with a ConvNet potential requires only a few hyper-parameters and no regularization.