BEL is a probabilistic description logic (DL) that extends the light-weight DL EL with a joint probability distribution over the axioms, expressed with the help of a Bayesian network (BN). In recent work it has been shown that the complexity of standard logical reasoning in BEL is the same as performing probabilistic inferences over the BN. In this paper we consider conjunctive query answering in BEL. We study the complexity of the three main problems associated to this setting: computing the probability of a query entailment, computing the most probable answers to a query, and computing the most probable context in which a query is entailed. In particular, we show that all these problems are tractable w.r.t. data and ontology complexity.
Ceylan, I., Penaloza, R. (2015). Probabilistic query answering in the Bayesian description logic BεL. In Scalable uncertainty management : 9th international conference, SUM 2015, Québec City, QC, Canada, September 16-18, 2015 ; proceedings (pp.21-35). Springer Verlag [10.1007/978-3-319-23540-0_2].
Probabilistic query answering in the Bayesian description logic BεL
Penaloza, R
2015
Abstract
BEL is a probabilistic description logic (DL) that extends the light-weight DL EL with a joint probability distribution over the axioms, expressed with the help of a Bayesian network (BN). In recent work it has been shown that the complexity of standard logical reasoning in BEL is the same as performing probabilistic inferences over the BN. In this paper we consider conjunctive query answering in BEL. We study the complexity of the three main problems associated to this setting: computing the probability of a query entailment, computing the most probable answers to a query, and computing the most probable context in which a query is entailed. In particular, we show that all these problems are tractable w.r.t. data and ontology complexity.File | Dimensione | Formato | |
---|---|---|---|
CePe-SUM15.pdf
accesso aperto
Tipologia di allegato:
Submitted Version (Pre-print)
Dimensione
359.42 kB
Formato
Adobe PDF
|
359.42 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.