First-Order Context-Specific Likelihood Weighting in Hybrid Probabilistic Logic Programs

Statistical relational AI and probabilistic logic programming have so far mostly focused on discrete models. The reasons for this is that one needs to provide constructs succinctly model the independencies in such models, also efficient inference.
 Three types of are important represent exploit scalable inference hybrid models: conditional elegantly modeled Bayesian networks, context-specific naturally represented by logical rules, amongst attributes related objects models expressed combining rules.
 This paper introduces a language, DC#, which integrates distributional clauses' syntax semantics principles programs. It represents three qualitatively. More importantly, we introduce algorithm FO-CS-LW DC#. first-order extension likelihood weighting (CS-LW), novel sampling method exploits ground upgrades CS-LW with unification rules case.