Technical Report No. 174 - Abstract

Kristian Kersting, Luc De Raedt
Basic Principles of Learning Bayesian Logic Programs

Bayesian logic programs tightly integrate definite logic programs with Bayesian networks in order to incorporate the notions of objects and relations into Bayesian networks. They establishing a one-to-one mapping between ground atoms and random variables, and between the {\it immediate consequence operator} and the {\it directly influences by} relation. In doing so, they nicely separate the qualitative (i.e. logical) component from the quantitative (i.e. the probabilistic) one providing a natural framework to describe general, probabilistic dependencies among sets of random variables. In this paper, we present results on combining {\it Inductive Logic Programming} with Bayesian networks to learn both the qualitative and the quantitative components of Bayesian logic programs from data. More precisely, we show how the qualitative components can be learned by combining the inductive logic programming setting {\it learning from interpretations} with score-based techniques for learning Bayesian networks. The estimation of the quantitative components is reduced to the corresponding problem of (dynamic) Bayesian networks.