Query Answering in Bayesian Description Logics

Abstract. The Bayesian Description Logic (BDL) BEL is a probabilistic DL, which extends the lightweight DL EL by defining a joint probability distribution over EL axioms with the help of a Bayesian network (BN). In the recent work, extensions of standard logical reasoning tasks in BEL are shown to be reducible to inferences in BNs. This work concentrates on a more general reasoning task, namely on conjunctive query answering in BEL where every query is associated to a probability leading to different reasoning problems. In particular, we study the probabilistic query entailment, top-k answers, and top-k con-texts as reasoning problems. Our complexity analysis suggests that all of these problems are tractable under certain assumptions.