AAAI Publications, Twenty-Third International FLAIRS Conference

Font Size: 
Using Constraint Satisfaction for Learning Hypotheses in Inductive Logic Programming
Roman Bartak, Ondrej Kuzelka, Filip Zelezny

Last modified: 2010-05-06

Abstract


Inductive logic programming (ILP) is a subfield of machine learning which uses first-order logic as a uniform representation for examples, background knowledge and hypotheses. In this paper we deal with the template consistency problem, i.e., the problem of finding a consistent hypothesis, which is essential in ILP. A consistent hypothesis entails all positive examples and no negative example. The paper suggests using constraint satisfaction techniques to find suitable unifications of variables in the template and to perform the incurred subsumption checks.

Full Text: PDF