Graphical Models for Discrete and Continuous Data

We introduce a general framework for undirected graphical models. It generalizes Gaussian graphical models to a wide range of continuous, discrete, and combinations of different types of data. We also show that the models in the framework, called exponential trace models, are amenable to efficient estimation and inference based on maximum likelihood. As a consequence, we expect applications to a large variety of multivariate data that have correlated coordinates.