Constraint models for reasoning on unification in inductive logic programming

Abstract

Inductive Logic Programming (ILP) deals with the problem of finding a hypothesis covering all positive examples and excluding negative examples. One of the sub-problems is specifying the structure of the hypothesis, that is, the choice of atoms and position of variables in the atoms. In this paper we suggest using constraint satisfaction to describe which variables are unified in the hypotheses. This corresponds to finding the position of variables in atoms. In particular, we present a constraint model with index variables accompanied by a Boolean model to strengthen inference and hence improve efficiency. The efficiency of models is demonstrated experimentally. © 2010 Springer-Verlag.