Graphical Models in Science

(3 Credit Hours) · Undergraduate

This course provides an overview of probabilistic graphical models and their applications to various scientific fields. Probabilistic graphical models allow us to illustrate the dependence structure of numerous random variables by means of a graph, where nodes represent the random variables and edges or arrows connect pairs of nodes to describe their dependence structure. Topics of study include: foundations of statistics, linear dependence, nonlinear dependence, conditional independence, information theory, undirected and directed graphical models, inference of graphical models, model selection, high-dimensionality, sparsity, optimization, and statistical reproducibility. These topics will be discussed alongside real scientific data applications, including neuroscience, genomics, astronomy, forensic science, psychology, and other fields.