ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Genetic discovery using ontology mappingofobservatio

To: "[ontolog-forum] " <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Michael Brunnbauer <brunni@xxxxxxxxxxxx>
Date: Mon, 13 May 2013 12:30:22 +0200
Message-id: <20130513103022.GA2244@xxxxxxxxxxxx>

Hello Rich,    (01)

IMO, if your system is supposed to come up with something that really can be 
called an ontology or an explanation of a phenomenon, it should have a search 
tree that is so big it will never produce anything useful.    (02)

Regards,    (03)

Michael Brunnbauer    (04)

On Sun, May 12, 2013 at 05:14:39PM -0700, Rich Cooper wrote:
> Thanks Michael,
> 
> My comments are below,
> -Rich
> 
> Sincerely,
> Rich Cooper
> EnglishLogicKernel.com
> Rich AT EnglishLogicKernel DOT com
> 9 4 9 \ 5 2 5 - 5 7 1 2
> 
>       Hello Rich,
> 
>       >Here is a recent patent I found a few
> minutes ago
>       >which to some degree could be applicable,
> though
>       >it is addressed to generating ontologies
> for
>       >business applications:
> 
>       If I understand this correctly, it
> describes a method of generating an
>       OWL ontology from an annotated XSD file.
> All terms in the generated ontology 
>       are already present in the XSD file or the
> annotations. XML files using that 
>       XSD can then be converted to triples using
> that ontology. 
> 
>       An approach like this enables you to query
> the data with SPARQL - which may
>       be a progress over what you could do
> before. But using reasoning or rules
>       for some advantage probably requires a
> much bigger investment of time.
> 
> Yes, this patent was simply one that shows
> generation of an ontology.  I am personally more
> interested in the methods of generating ontology
> from data than the specific use of OWL/XSD/SPARQL,
> but those techniques are useful for the purpose of
> seeing how these XML technologies are available.  
> 
> For example, the EHR technology is based on XML
> structuring, and EHRs are required to meet the
> specs on the CCR (continuing care record) and CCD
> (continuing care document) so that there are
> standard XML forms of the EHR.  That means the EHR
> information can be relied upon to provide SOAP
> (subjective, objective, assessment and plan)
> medical record in an XML supported document for
> each patient.  
>       
>       > By that description, I am envisioning a
> system
>       > that searches (think And/Or graph
> search) for the
>       > best explanation in a logical
> combination of
>       > evidence fragments.
> 
>       Something like this ? 
> 
>       
> http://www.aaai.org/Papers/Workshops/2006/WS-06-08
> /WS06-08-010.pdf
> 
> Thanks, that is a very interesting paper which I
> have just put on my reading list.  I like the way
> they flexibly shift between belief systems and
> search cones as explanations of the And/Or
> solution forest.  
>       
>       > So a discovery system that browses
> through all
>       > that data could put together logical
> combinations
>       > of evidence from each case, and produce
> millions
>       > of explanations as logical combinations
> of
>       > evidence terminals.  Its that "logical
>       > combinations of evidence terminals" that
> I am
>       > calling (perhaps inelegantly) an
> ontology.  
> 
>       What do you do with the millions of
> "explanations" ? If you do not reason with
>       them, I would not see them as part of an
> ontology. Isn't the system supposed to
>       yield the best "explanation" instead of
> millions ?
> 
> I was referring to creating an ontology using
> automated methods.  That means each ontology
> explanation of the database of phenomenon stands
> as an explanation of the phenomenon.  If there are
> millions of such explanatory ontologies, there is
> value in finding the ones that appear consistently
> in explanatory ontologies that also have specific
> other explanations.  That consistent pairing of
> explanatory chunks can help users understand how
> the pairing works, perhaps leading to new science
> compared to doing without that pairing.  
>       
>       > The reason I call it an ontology, even
> though it
>       > was not put together by humans, is
> because it
>       > represents the tightest explanation of
> the
>       > evidence in Occam's sense.
> 
>       I would say what you mean is the smallest
> set of ground facts with a 
>       correlation to a given desease. If you
> mean this, you can hardly call it
>       an ontology.
> 
> I don't mean that.  I mean the explanatory power
> of automatically generated ontologies, supported
> by ghastly large amounts of evidence, provides the
> foundation for understanding phenomenon that are
> far more complex than humans can generate on our
> own.  We would be more focused on the
> understanding of such pairings of ontology phrases
> than we could be otherwise; we would not have to
> use human hours to produce potentially useful
> leads, but instead be able to use the obsessive
> power of the computing system to generate more
> focused leads.  
>       
>       > I am suggesting that perhaps we should
> jettison the
>       > human generation of the ontology and
> substitute a
>       > method for automatic generation of the
> ontology
>       > without concern for whether a human
> understands
>       > the reasoning at any intuitive level.
> 
>       I do not see any real reasoning in the
> system you describe.
> 
> I didn't intend to claim reasoning in the system,
> other than the logical combination of
> explanations.  The reasoning is in the humans, who
> would have less overwhelming databases, instead
> using the generated explanation logic, i.e., the
> automatically generated ontologies.  
>       
>       Regards,
> 
>       Michael Brunnbauer
> 
> Thanks for your thoughts, and especially the
> reference you posted,
> 
> -Rich
>      (05)


>  
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/ 
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
>      (06)

-- 
++  Michael Brunnbauer
++  netEstate GmbH
++  Geisenhausener Straße 11a
++  81379 München
++  Tel +49 89 32 19 77 80
++  Fax +49 89 32 19 77 89 
++  E-Mail brunni@xxxxxxxxxxxx
++  http://www.netestate.de/
++
++  Sitz: München, HRB Nr.142452 (Handelsregister B München)
++  USt-IdNr. DE221033342
++  Geschäftsführer: Michael Brunnbauer, Franz Brunnbauer
++  Prokurist: Dipl. Kfm. (Univ.) Markus Hendel    (07)

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (08)

<Prev in Thread] Current Thread [Next in Thread>