A Probabilistic graphical model is a model where a graph structure is used to represent random variables together with their conditional dependencies. Most commonly used models are Bayesian networks (directed models) and Markov random fields (undirected models). Graphical models provide a useful framework for representing domains with complex interactions, with applications in computer vision, natural language processing, medical diagnosis, computational biology, and so on. 

keywords: Bayesian Network, Gibbs distribution, Markov Random fields, Influence diagram