Learning logic programs through divide, constrain, and conquer

TL;DR

This paper presents a novel inductive logic programming method that integrates divide-and-conquer strategies with constraint-driven search, enabling efficient learning of complex, recursive logic programs with predicate invention across multiple domains.

Contribution

It introduces an anytime, scalable approach that combines classical divide-and-conquer with modern constraints, supporting predicate invention and improving learning efficiency.

Findings

Increased predictive accuracies across domains

Reduced learning times compared to existing methods

Able to learn optimal, recursive, and large programs

Abstract

We introduce an inductive logic programming approach that combines classical divide-and-conquer search with modern constraint-driven search. Our anytime approach can learn optimal, recursive, and large programs and supports predicate invention. Our experiments on three domains (classification, inductive general game playing, and program synthesis) show that our approach can increase predictive accuracies and reduce learning times.