Describir: Learning Structure and Parameters in Bayesian Networks Using an Implicit Framework