Generalized subsumption and its applications to induction and redundancy (Q1110345)

A theoretical framework and algorithms for the study of induction of definite clauses are introduced. This forms a strong model called generalized subsumption on which suitable search space can be built. It highlights some of the properties a generalization model should have. Applications of the theory and implications to AI are discussed such as methods to detect redundant clauses in logic programs and redundant atoms within clauses, finding of most specific generalization, interaction between generalization, induction, relevance, redundancy and structure. An improvement of the most general refinement operator is suggested.