John,
Concerning your proposed revision: (01)
[John Bateman] Further possible revision?... (02)
(2) Each of the existing upper ontologies differs in specifics of
implementation, but we all agree that use of a formally defined
common library of upper ontologies with explicit interrelations is
the most satisfactory method for achieving semantic interoperability
that can scale from a few applications to large communities of diverse
knowledge-based reasoning systems. (03)
[PC] The big problem I perceive with this phrasing is that, though it
eliminates completely any possible future where one upper ontology may
eventually become the dominant one for machine semantic communication
(rather like the Windows Operating System for programs or the English
language for international scientific conferences), it also leaves wide
open the interpretation that there is no reason whatever to limit the
number of upper ontologies, interrelations are easy to find, and maybe
even more is better, like books in a library. For "scalable" semantic
interoperability, a CUO is (IMO) more than just "most satisfactory", it
can be readily projected to be the most "cost effective" by considering
alternative scenarios involving dozens of separately developed
ontology-driven reasoning systems. Consider how to achieve **the same
accuracy** in transmission of conceptual information among multiple
independently developed reasoning systems without a CUO -- if possible
at all, it must be much more complicated than a CUO. (04)
To reiterate the phrasing at issue:
> (2) Each of the existing upper ontologies differs in specifics of
> implementation, but we all agree that use of some formally defined
> common upper ontology is the most cost-effective method for achieving
> semantic interoperability that can scale from a few applications to
> large communities of diverse knowledge-based reasoning systems. (05)
[PC] I can't think of any reason that anyone would interpret that
statement as meaning that this effort is aimed at creating "the one
accepted theory of everything", unless one goes back into the archives
of the IEEE-SUO discussions and discovers that everyone on that list
denounces the possibility of creating some "one accepted theory of
everything", and from that comes away with the notion that someone,
somewhere must have proposed such a theory, and maybe this crowd is
what everyone else is denouncing. I don't think the audience to which
this communique is addressed is at risk of drawing such a conclusion.
For one thing, it is clear that the intention that comes through from
the whole statement is that each of the existing upper ontologies will
continue to have its own identity. For another, the use of the word
"communities" in this sentence should imply quite clearly that there
may be more than one such community, each using some (but not
necessarily the same) CUO. (06)
[PC] Since any statement in language that does not go on for pages with
clarifications is going to give rise to concerns about
misinterpretation, I would suggest that we put all these detailed
clarifications and caveats in a Wiki page which will be pointed to in
the communique as "Extended Remarks" that can contain warnings about
all of the potential misinterpretations that we don't want to convey.
One such caveat could be that the panelists do not expect any agreement
in the near future on one fully axiomatized upper ontology that will
satisfy all potential users. That is implied in the statement, and
can be made explicit in the Extension Wiki without giving a false
impression that there is little real value beyond some "satisfaction"
in using a common upper ontology for interoperability. (07)
I have created an "Extended Remarks" Wiki: (08)
http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit/ExtendedRem
arks
. . . to contain such clarifications and extended comments, and have
placed in it remarks that I believe address (at least some of) your
concerns and those of Bill Andersen. Feel free to modify or add to
them. The Joint Communique will be whatever the panelists decide it
should be, but I think that a vision statement pointing toward the
future, based on projections from past experience, would be most
appropriate for a meeting of this type. I think that in a public
statement, succinctness is a great virtue, rather like an "Executive
Summary", and a proper division of labor between the Communique and the
Extended Remarks would keep the communique short to express the views
of the panelists in words that are accurate in the sense intended,
without being so diluted by caveats as to leave an impression that
there is no substance. The extended remarks can contain all the
clarifications needed to avoid possible misinterpretations by those who
wonder about possible alternative interpretations they may perceive.
If the panelists agree, they and the key participants may feel free to
address potential misinterpretations and provide more clarifications in
the Extended Remarks, which do not have to be succinct or quotable.
The Extended Remarks page is freely editable. (09)
Since this year marks the 50th anniversary of "Artificial Intelligence"
(yes, 50) - it may be appropriate to look back at the proposal (see
http://www-formal.stanford.edu/jmc/history/dartmouth/dartmouth.html)
for the Dartmouth Conference which popularized that term. The second
sentence of the proposal stated (and has been quoted in retrospective
articles): (010)
"The study is to proceed on the basis of the conjecture that every
aspect of learning or any other feature of intelligence can in
principle be so precisely described that a machine can be made to
simulate it." (011)
The Joint Communique might also serve as a (more modest) succinct
vision of what can be reasonably expected to be accomplished by
interrelating the existing upper ontologies -- which would be much
narrower than the goals of the Dartmouth group. The Extended Remarks
might contain the more down-to-earth qualifiers as to what we know can
already be done with existing systems, and caveats about possible
problems in the future. (012)
Pat (013)
Patrick Cassidy
MITRE Corporation
260 Industrial Way
Eatontown, NJ 07724
Mail Stop: MNJE
Phone: 732-578-6340
Cell: 908-565-4053
Fax: 732-578-6012
Email: pcassidy@xxxxxxxxx (014)
-----Original Message-----
From: uos-convene-bounces@xxxxxxxxxxxxxxxx
[mailto:uos-convene-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John A.
Bateman
Sent: Sunday, March 05, 2006 1:38 PM
To: Upper Ontology Summit convention
Subject: Re: [uos-convene] "common" upper ontology? (015)
Further possible revision?... (016)
(2) Each of the existing upper ontologies differs in specifics of
implementation, but we all agree that use of a formally defined
common library of upper ontologies with explicit interrelations is
the most satisfactory method for achieving
semantic interoperability that can scale from a few applications to
large communities of diverse knowledge-based reasoning systems. (017)
Reasons: (018)
(i) while I agree that the debate between 'standard'/'common' upper
ontology and lattices is rather internal, the repercussions
are not. As soon as the notion of a 'common ontology' is put
forward, we can draw on decades of experience of not having
achieved one. To suggest, even indirectly by leaving open
possible interpretations, that this
group thinks that it is going to come
up with the one accepted theory of everything is not going
to sound serious. Once this interpretation is in place, it
is more difficult to backtrack and say, oh, actually we meant
something more sophisticated.... (019)
(ii) I would not like to be asked to prove that our approach
(I mean in Bremen) is maximally "cost-effective" at this point
in time! Depends precisely on the tasks, the level of
interoperability targetted, the results that will be reached
over the next 10 years with establishing common libraries, and and. (020)
John B.
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uos-convene/
To Post: mailto:uos-convene@xxxxxxxxxxxxxxxx
Community Portal: http://ontolog.cim3.net/
Shared Files:
http://ontolog.cim3.net/file/work/UpperOntologySummit/uos-convene/
Community Wiki:
http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uos-convene/
To Post: mailto:uos-convene@xxxxxxxxxxxxxxxx
Community Portal: http://ontolog.cim3.net/
Shared Files: http://ontolog.cim3.net/file/work/UpperOntologySummit/uos-convene/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit (021)
|