Inverse Entailment in Nonmonotonic Logic Programs
Proceedings of the 10th International Conference on
Inductive Logic Programming (ILP-2000),
Lecture Notes in Artificial Intelligence 1866,
Springer-Verlag, pages 209-224, 2000.
Inverse entailment (IE) is known as a technique for finding
inductive hypotheses in Horn theories.
When a background theory is nonmonotonic, however,
IE is not applicable in its present form. The purpose of this paper is
extending the IE technique to nonmonotonic inductive logic programming (ILP).
To this end, we first establish a new entailment theorem in normal logic
programs, then introduce the notion of contrapositive programs.
Finally, a theory of IE in nonmonotonic ILP is constructed.
Full Paper (gzipped postscript 174K)
Slide (pdf 111K)