[SEEK-Taxon] SEEK-Taxon Dave Thau Ontologies Update

[Beach, James H]

SEEK Taxon Members -- 

Here is an update from Dave Thau about his work with ontologies for the
SEEK-Taxon group since our meeting in Lisbon.  Dave will be at our
meeting in Santa Barbara next week. Let's discuss Dave's findings with
him then. 

-- Jim

 
--------------------------------
James H. Beach
Biodiversity Research Center
University of Kansas
1345 Jayhawk Boulevard
Lawrence, KS 66045, USA
Tel: 785 864-4645, Fax: 785 864-5335
Televideocon: (H.323): 129.237.201.102


-----Original Message-----
From: thau at learningsite.com [mailto:thau at learningsite.com] 
Sent: 16 January, 2004 4:58 PM
To: Beach, James H
Subject: update

Hi Jim,

Here's an update on my progress since our meeting a month ago.

I've been creating a workable OWL DL instantiation of Jesse Kennedy's
schema.  OWL DL is the version of OWL which lets you use the reasoners
being developed by the semantic web community.

Under the constraints of OWL DL, there are two main approaches to
creating an ontology based on Jesse's schema.  One is to consider taxa
as individuals, the other to consider taxa as collections of specimens
(or character states, or something else).

The route to take depends on how easily they can answer the questions
we'll want to ask the ontology, and how they deal with scaling issues.
It's probably the case that each route makes sense under some set of
circumstances.

I've been pursuing the first route and have an OWL DL ontology which
works with one of the standard reasoners (pellet) to provide consistency
checking and reasoning.  You can feed the ontology and a set of taxa to
the reasoner and ask if the taxa violate the ontology's rules (e.g. a
taxon has 2 parents, or something his is stated to be a phylum has a
species as a parent).  You can also ask for the children of a taxon, and
if a taxon is a junior synonym of another one, it can give you the
children of the accepted taxon.  

I've attached the current ontology here, along with a small data set.  
You won't be able to do anything with them without a reasoner, but at
least you can look at them and see what OWL looks like.  I'll
demonstrate them working when I see you two Mondays from now.

The next steps here are to generate bigger data sets and see how they
scale, and then to work up an ontology following the second route.  Now
that I have the first one working to some extent, the second one will
come much more quickly.

I've also sketched out the beginnings of a framework for a distributed
ontology creation environment.  This framework has two goals: to put
ontology creation and maintanence in a grid-like environment, and to
allow data providers to easily provide and update their data while
avoiding some of the pitfalls of a centralized system.  I'm writing
these ideas up and will have something for folks to read in Santa
Barbara.

I pursued this framework to provide an idea of how OWL might be useful
to the taxon group when doing taxonomic work in a distributed context.

Finally, I've been helping Aimee out a TINY bit with an xslt document
which converts DiGIR responses to EML documents.  I love xslt!

That pretty much covers it.  Happy New Year!

Dave.





-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mercury.nceas.ucsb.edu/ecoinformatics/pipermail/seek-taxon/attachments/20040120/52cd72c0/attachment.htm
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: scaling_abox.owl
Url: http://mercury.nceas.ucsb.edu/ecoinformatics/pipermail/seek-taxon/attachments/20040120/52cd72c0/scaling_abox.bat
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: 9_abox_dataset.owl
Url: http://mercury.nceas.ucsb.edu/ecoinformatics/pipermail/seek-taxon/attachments/20040120/52cd72c0/9_abox_dataset.bat