Re: Your RuleML presentation (ACTION on hak)

Let me make a very simple example.

Say you have three OWL (more precisely Horn-SHIQ) statements:

C \sqcap D \equiv \bot (i.e. C,D are disjoint)
B \sqsubseteq C (i.e. B is a subclass of C)

and the ABox statement

a\in B (i.e. individual a is a member of B).

Then this could be translated into logic programming as

inc <- C(x), D(x)
C(x) <- B(x)
B(a)

[I'm not sure the KAON2 algorithms would yield exactly this, but it 
serves as an intuitive example]

For the derivation of positive ground facts, like C(a) in this example, 
the first rule (integrity constraint) can be ignored.

If, however, the original knowledge base is inconsistent, e.g. by also 
containing a\in D, then we would have another fact D(a) in the logic 
program, which means that ?- inc. would be derivable, showing the 
inconsistency.

Basically, the integrity constraints derived from Horn-SHIQ using the 
KAON2 algorithms can be ignored, if the original knowledge base is 
consistent - because they will never fire.

Note that this only works for Horn-SHIQ, which is a superset of DLP. It 
won't work for larger fragments of OWL DL in general.

The paper is here: 
http://www.aifb.uni-karlsruhe.de/WBS/phi/resources/publications/owlvialp06.pdf
It also contains the (extensive) example I showed on one of the slides 
(page 8 top).

Best Regards,

Pascal.


Hassan Aït-Kaci schrieb:
> Hi Pascal,
> 
> I mentioned to the RIF WG your 'inc' predicate that you use for
> capturing DL features such as class mutual exclusion and negation
> in general. From what I understood in your answer to my question
> (How the inconsistent predicate 'inc' is used by the interpret of
> the LP translation?), you said that it was ignored, and therefore
> such inconsistensies were tolerated in the translation.
> 
> The WG wishes to have examples for such things as ignoring features
> (such as negation). Can you provied yours as one?
> 
> Thanks.
> 
> -hak

-- 
Dr. habil. Pascal Hitzler
Institute AIFB, University of Karlsruhe, 76128 Karlsruhe
email: hitzler@aifb.uni-karlsruhe.de    fax: +49 721 608 6580
web:   http://www.pascal-hitzler.de   phone: +49 721 608 4751
        http://www.neural-symbolic.org

Received on Wednesday, 15 November 2006 05:17:58 UTC