Efficient Methods for Unsupervised Learning of Probabilistic Models

Abstract

In this thesis I develop a variety of techniques to train, evaluate, and sample from intractable and high dimensional probabilistic models. Abstract exceeds arXiv space limitations -- see PDF.