Stochastic Logic Programs: Sampling, Inference and Applications

Algorithms for exact and approximate inferencein stochastic logic programs (SLPs) are presented, based respectively, on variable elimination and importance sampling. We then showhow SLPs can be used to represent prior distributions for machine learning, using(i) logic programs and(ii) Bayes net structures as examples.Drawing on existing work in statistics, we applytheMetropolis-Hasting algorithm to construct aMarkov chain which samples from the posteriordistribution. A Prolog implementation for this isdescribed. We also discuss the possibility of constructing explicit representations of the posterior