Statistical Machine Learning in Markov Random Fields

Generative Model for Extracting the Data Expression

Abstract

In this article, we introduce the foundations of the generative model approach in statistical machine learning theory based on Markov random fields. The purpose of the generative model approach is to create a probabilistic model that mimics the source from which the observed data were generated. After explaining Markov random fields and the generative model approach, we introduce the restricted Boltzmann machine, which is a basic model of a deep generative model.