MC-TopLog


MC-TopLog is an ILP system that is sound and entailment-complete for hypothesis finding. It has another two key features that destinguish it from other multi-clause learners: (1) co-generalization (generalise multiple examples together); (2) uses a top theory as declarative bias.

The first version of MC-TopLog implemented the algorithm of Top-directed Theory Derivation (TDTD). Details of TDTD can be found in Dianhuan Lin's master thesis 'Efficient, Complete and Declarative Search in Inductive Logic Programming'

Later, TDTD is extended to TDTcD (Top-directed Theory co-Derivation) based on the idea of constructing common generalisations. TDTcD is described in the paper 'MC-TopLog: Complete Multi-clause Learning Guided by a Top Theory', which is going to be published in the proceedings of ILP11. Materials and ILP systems used in the experiments of this paper are here.

MC-TopLog has been applied to two real-world applications. Initial results are described in 'Does Multi-clause Learning Help in Real-world Applications?', which is also going to be published in the proceedings of ILP11. Materials and ILP systems used in the experiments of this paper are here.