Adventures of Two Little OWLs in Rule Land
Adventures of Two Little OWLs in Rule Land
Abstract. Combining ontological and rule-based modelling can be an onerous task, from the choice of a suitable semantic framework (there are quite a few) to the selection of a chain of tools for supporting it (there are just a few). Typical solutions combine not only the advantages but also the difficulties of both domains, especially regarding computational complexity. For the recently introduced light-weight profiles of OWL 2, however, the situation is remarkably different. Here we find that existing rule-based systems can rather easily be adopted to support ontological inferencing using established algorithmic methods. This is well-known for OWL RL – “RL” is for “Rule Language” after all – but much less so for OWL EL.
In this talk, we take a closer look at this exciting grey area between light-weight ontologies and rules where both approaches are close enough to allow for an easy combination. We recall the features of OWL EL and RL, and explain how reasoning tasks in both languages can be answered by common rule systems with only a slight transformation of syntax. This approach uses rules as a computational formalism for implementing OWL reasoning without implying a semantic connection: even production rule systems could be used. Going further, we aim at a more intimate semantic combination of (logical) rules, OWL EL, and OWL RL, carefully tuned to allow efficient implementation in polynomial time. Further insights into matters of practical efficiency are gained from recent results on the worst-case space requirements of OWL EL inferencing, and from our experiences with the prototype implementation Orel.
Published at 1st International Workshop on Business Models, Business Rules and Ontologies (BuRo 2010) (Talk)
Download PDF (last update: 21 Sep 2010)
Citation details
- Markus Krötzsch. Adventures of Two Little OWLs in Rule Land. 2010.
Remarks
This invited talk included results that were published in the paper Efficient Inferencing for OWL EL. Further references are given at the end of the slides.
