OASIcs.ICLP.2017.3.pdf
- Filesize: 409 kB
- 5 pages
In this paper we empower the ontology-based query answering framework with the ability to reason on the properties of “known” (non-anonymous) and anonymous individuals. To this end, we extend Datalog+/- with epistemic variables that range over “known” individuals only. The resulting framework, called datalog^{\exists,K}, offers good and novel knowledge representation capabilities, allowing for reasoning even on the anonymity of individuals. To guarantee effective computability, we define shyK, a decidable subclass of datalog^{\exists,K}, that fully generalizes (plain) Datalog, enhancing its knowledge modeling features without any computational overhead: OBQA for shyK keeps exactly the same (data and combined) complexity as for Datalog. To measure the expressiveness of shyK, we borrow the notion of uniform equivalence from answer set programming, and show that shyK is strictly more expressive than the DL ELH. Interestingly, shyK keeps a lower complexity, compared to other Datalog+/- languages that can express this DL.
Feedback for Dagstuhl Publishing