Bayesian Networks are models that use a directed acyclic graph (DAG) to represent a set of variables and their conditional dependencies. Each node in the graph represents a variable, and edges denote conditional dependencies between these variables. They facilitate decision-making processes by enabling reasoning under uncertainty. For instance, in medical diagnosis, Bayesian Networks can model the probabilistic relationships between diseases and symptoms, helping infer the likelihood of a disease given observed symptoms.