Inductive Logic Programming: From Machine Learning to Software Engineering

Although Inductive Logic Programming (ILP) is generally thought of as a research areaat the intersection of machine learning and computational logic, Bergadano and Gunetti propose thatmost of the research in ILP has in fact come from machine learning, particularly in the evolution ofinductive reasoning from pattern recognition, through initial approaches to symbolic machinelearning, to recent techniques for learning relational concepts. In this book they provide anextended, up-to-date survey of ILP, emphasizing methods and systems suitable for softwareengineering applications, including inductive program development, testing, andmaintenance.Inductive Logic Programming includes a definition of the basic ILP problem and itsvariations (incremental, with queries, for multiple predicates and predicate inventioncapabilities), a description of bottom-up operators and techniques (such as least generalgeneralization, inverse resolution, and inverse implication), an analysis of top-down methods(mainly MIS and FOIL-like systems), and a survey of methods and languages for specifying inductivebias.Logic Programming series