For example,
Localized engineer may input "football_player \exists hasTeam. Team", this
is easy, you can capture it as A\exists R.B. But the engineer may input
much more complex axioms such that pattern discovery cannot find them. The
motivation is that the user should have a better observation of what kind
of axioms to be discovered.
Actually, it's better if you allow "natural language descriptions of
classes as input". :)
How far have you been doing this? Maybe a gsoc project?
Best.
On Tue, Mar 11, 2014 at 6:22 AM, Jens Lehmann <
[email protected]> wrote:
>
> Hello,
>
> Am 10.03.2014 20:39, schrieb Cong Wang:
> >
> > Also, to enrich ontology axiom, we may ask user to input some key axiom
> > regarding the topic. Since in [1], the pattern is actually generic, it
> > has no semantics regarding a topic. Like, the pattern are just A \sub B,
> > A \sub \exists R.B, etc. Then we use such axiom type to learn axioms.
> > However, i believe it can be better if we have some guideline.
>
> Can you explain this part a bit more? Do you mean that the user should
> be able to manually extend/revise the patterns?
>
> By the way, we are now trying to go beyond reference [1] you mentioned
> by also including NLP in the learning process (using both the RDF data
> and natural language descriptions of classes as input for learning
> whereas previously only one of the those has been used).
>
> Kind regards,
>
> Jens
>
> --
> Dr. Jens Lehmann
> AKSW Group, Department of Computer Science, University of Leipzig
> Homepage: http://www.jens-lehmann.org
> GPG Key: http://jens-lehmann.org/jens_lehmann.asc
>
--
Cong Wang
Ph.D Candidate,
Kno.e.sis Center,
Wright State University.
------------------------------------------------------------------------------
Learn Graph Databases - Download FREE O'Reilly Book
"Graph Databases" is the definitive new guide to graph databases and their
applications. Written by three acclaimed leaders in the field,
this first edition is now available. Download your free book today!
http://p.sf.net/sfu/13534_NeoTech
_______________________________________________
Dbpedia-gsoc mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbpedia-gsoc
