[seek-taxon] RE: some ideas

Sun Sep 14 16:04:14 PDT 2003

Howdy,

Dave Thau seems like a bright guy that we might benefit from.  However, it
seems like he would be working in isolation from our previous work.  We
are already working a bit too much as fragmented groups rather than as a
coherent team.  To add another fragment would likely not be productive,
but rather just distract us.  Is there a convenient way to bring him up to
speed on the various things members of our group have done and then
interact with him regularly during the contract period so that he can
build on our our past work and help us to converge on a common vision?

Bob

On Sat, 13 Sep 2003, Beach, James H wrote:

>
> Dave --
>
> Thanks for this interesting set of ideas.  Let me forward them to Matt
> Jones, Jessie Kennedy, Susan Gauch,  and Bertram Ludaescher to see what they
> think.  I will (and maybe they will hopefully also) get back to you early
> next week and we'll make some decisions.
>
>
> Jessie, Bertram, Susan and Matt --  Would you have any thoughts on these
> suggestions from Dave Thau on possible work areas for semantic processing of
> classifications.  These are just early thoughts from Dave, he has not talked
> in depth with anyone of us yet.
>
> We have proposed to Dave that we would hire him as a consultant on the SEEK
> Taxon group for an initial 30 day period to research the scope of challenges
> with the semantic processing of multiple classifications.  So this is just
> very preliminary and it would be very useful for SEEK if you had any
> thoughts for Dave on the challenges he describes.
>
> -- 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: 11 September, 2003 6:46 PM
> > To: Vieglais, David A; Beach, James H
> > Subject: some ideas
> >
> >
> > Howdy!
> >
> > I'd like to try out some SEEK ideas just to make sure I'm
> > thinking correctly.  Let me know if any of these sound
> > particularly interesting or particularly irrelevant.  Let me
> > know if I'm totally off base here too.
> > I haven't yet scheduled any meetings or phone calls with
> > anyone, so this
> > comes out of a bit of a vacuum.
> >
> > I've come up with four basic topics lurking around the TG/SMS
> > border so
> > far:
> >
> > 1.  Using SMS representations and languages to reason over
> > multiple taxonomic hierarchies.
> >
> > 2.  Using SMS to aid in building an index, registry,
> > collection of assertions, whatever it is that the TG will be serving.
> >
> > 3.  Figuring out what to do with data sets describing taxa at
> > different levels (e.g. genus level data vs. species level data).
> >
> > 4.  Figuring out what to do with data sets containing
> > uncertain taxonomic information.
> >
> > Here are some more details about each of these.  Read on, or
> > if it's too much for an email, let me know if a phone call
> > would be more appropriate.
> >
> >
> > 1.  The SMS folks have various ontology representations and
> > inference engines.  It may be a good idea for the taxon group
> > to use these representations and engines when doing things
> > like reasoning over multiple conflicting hierarcies, or
> > simply responding to SMS queries.  There are a number of
> > things I could look into regarding that.  Some quick ideas:
> >   a.  working on a way for new taxonomic concept providers to
> > share data
> > with SEEK
> >   b.  helping translate the rules developed by the TG into
> > more SMS-like representations (like F-Logic and OWL)
> >   c.  using the SMS stuff to navigate conflicting hierarchies
> > - one thing to point out is that I already have a database of
> > conflicting taxonomic hierarchies buried in the
> > www.speciestoolkit.org.  Hidden in there are full > taxonomies
> > from species 2000, ITIS, Genbank, and a couple others.
> >
> > 2.  Using SMS to support feature extraction techniques that
> > may be useful in determining taxonomic concepts in EML data
> > sets and other types of documents. I'm a bit unclear about
> > what data are going to reside in a "taxanomic name server"
> > and what data are going to be calculated over a distributed
> > fashion, but assuming there's going to be some sort of index
> > of names or collection of assertions, or registry, SMS could
> > be a help in growing that index, collection or registry.
> >
> > 3.  The simple cases of SMS and TG working together are
> >   a.  SMS asks TG which of a set of names represent the same critters
> >   b.  SMS gives TG as set of names and wants the accepted names back
> >   c.  SMS wants a list of possible synonyms for some name
> >
> > However, these largely act within one level of a hierarchy.
> > I have some species names, I want the 'official' species
> > names.  What happens when SMS wants to reason over data sets
> > which act at different levels?  One data set may cover a
> > subset of a genus (e.g. Arizona desert Pheidoles) while
> > another might discuss a specific species (Pheidole titanus).
> > What role can the TG play in helping the SMS reason
> > accurately in cases like this?
> >
> > 4.  There seems to be a bit of confusion about what to do
> > with data sets which are unclear about the species they're
> > representing.  Those data sets may be rejected from analysis,
> > or perhaps the taxon group can help add a little certainty.
> > If a set of observations uses a common name, it may be
> > possible to narrow down which species they might mean using a
> > combination of information from that data set and some input
> > from the person contributing the set.  This would definitely
> > help the SMS in its quest to find data to contribute to an
> > analysis, but information generated like this should be
> > tagged with a certainty value of some sort. This whole area
> > is pretty rich for investigation.
> >
> > Ok... that's it for now.  Looking forward to your reactions.
> >
> > thau!
> >
> >
> >
> _______________________________________________
> seek-taxon mailing list
> seek-taxon at ecoinformatics.org
> http://www.ecoinformatics.org/mailman/listinfo/seek-taxon
>

 ====================================================================

 Robert K. Peet, Professor              Phone:  919-962-6942
 Department of Biology, CB#3280         Fax:    919-962-6930
 University of North Carolina           Cell:   919-368-4971
 Chapel Hill, NC  27599-3280  USA       Email:  peet at unc.edu

             http://www.bio.unc.edu/faculty/peet/

 ====================================================================