We give an overview of the relational Bayesian network modeling language. First the semantic concept of a random relational structure model is introduced, and then it is shown how such models can be represented with relational Bayesian networks. We consider a number of inference problems for relational Bayesian networks that range from elementary probabilistic queries to the computation of limit probabilities and learning problems. For some of these inference problems fully developed solution algorithms are available, for others we describe solution strategies by reduction to well-established logical inference and numerical optimization problems.
Get the full paper