Markov Random Fields (MRFs) are a type of probabilistic graphical model that represent the joint distribution of a set of random variables having a Markov property described by an undirected graph. They are particularly useful for modeling spatial dependencies and are widely used in image processing, computer vision, and statistical physics.