Current machine learning systems are often distinguished on the basis of their representation, which can either be propositional or first order logic. Systems belonging to the first category are often called attribute value learners, systems of the second category are called relational learners or inductive logic programming systems.