|
At 10:14 AM +0700 1/30/08, paola.dimaio@xxxxxxxxx wrote:
Can Cyc be redesigned-reused?
I mean an ontology is made of
knowledge (incl axioms)
represenation
formalization
what part of Cyc makes it useless?
It is not useless. It is probably the most highly used large
ontology ever constructed, and is the only one that is a commercial
success. OpenCyc, the free subset of Cyc, is used as a resource by
hundreds of academic and commercial projects. Cyc is also many orders
of magnitude larger than any other ontology.
Although some early grandiose claims for the Cyc project (which
were not taken seriously by anyone in the AI field) have not been
bourne out, this does not invalidate its power or utility as an
ontology.
What would be the best way to fix
that?
Cyc now represents approximately 200 man-years of effort. No
other ontology or KR project comes even close to this scale.
"Fixing" it, whatever that might mean, would require a
similar scale of effort. No US government agency will subsidize such
an effort, and I doubt if one could get EEC funding for it
either.
Pat
pdm
On Jan 30, 2008 6:10 AM, Avril Styrman <Avril.Styrman@xxxxxxxxxxx>
wrote:
John and Azamat,
comments below.
Lainaus "John F. Sowa" <sowa@xxxxxxxxxxx>:
> Azamat:
> > What matters is the whole new idea of subjecting
metaphysical
> > systems to axiomatization, the rigorous and systematic
analysis
> > of a system from precise definitions, axioms and rules,
what
> > Spinoza essayed in his philosophy.
>
> That is indeed an interesting idea. In mathematics, it has
proved
> to be very valuable. But in the empirical sciences, even
physics,
> it has had mixed results. Newton's achievements came from
applying
> mathematical techniques to ideas that had been developed (with
some
> use of mathematics) by Galileo and Kepler.
> As far as formalizing metaphysics, it is safe to say
that there are
> *no* major insights that have resulted from the formalization.
It
> is true that formal axioms are important for computer
applications,
> since computers, by themselves, have no intuitions whatever.
But
> the real insights have come from people whose intuitions were
at
> the level of Galileo, Faraday, and Einstein. There are no
formal
> achievements that are remotely comparable to the work of
Newton,
> Maxwell, or Minkowski. (But there are insights into
metaphysics
> that result from mathematical studies in physics, but the
insights
> originated in physics, not formal philosophy.)
>
> In summary, the attempts by Spinoza and Descartes to
formalize
> metaphysics were interesting failures. Even as late as the
20th
> century, when modern logic became available, attempts such as
> Carnap's Logische Aufbau were also interesting failures.
Carnap
> was a good logician with a strong background in physics and
> mathematics. But his attempt was a dead end. Nelson
Goodman
> made another interesting attempt, which developed some useful
> mathematics, but no new insights into metaphysics.
>
>
> Those of us who are working with logic and ontology
hope that
> some kind of formalization will be useful for major
applications.
> But so far, there have been *no* new insights into metaphysics
that
> have come from the process of axiomatizing Cyc or any other
formal
> system. On the contrary, the real insights have come from
the same
> source as all the other insights since antiquity: dedicated
study,
> observation, intuition, and discussion with teachers, students,
and
> colleagues. The insights from formalization, if any, were
modest
> at best.
We have to separate two things:
1) formalization
2) axiomatization
To axiomatize something, does not mean that we have to impose
a horrible string of mathematical formulas upon it. In the
case of philosophical ontology (the best part of metaphysics),
plain text will do. Of course, some principles such as
the
identity of indiscernibles and the indiscernibility of
identicals can be given in both ways, in text and in formulas,
and it helps understanding them. The act of axiomatization should
help understanding the subject, not to make it harder. Of course,
pictures and graphs help too.
To be clear, to state something axiomatically is clear, simply
because othervise it could be unclear. Over-formalization is a
threat that has to be avoided. The clearest cases of
over-formalization are the those where the thing that is
to be formalized, is actually required to understand the
formalization itself! Examples of these are e.g Frege-Russell,
von Neumann, and Zermelo -definitions of natural numbers.
Avril
PS:
> The largest of all attempts was the Cyc project,
which many people
> in AI regard as a failure. Cyc has had some useful
applications,
> but none of them have been sufficiently successful to pay for
the
> many millions of dollars that were invested in the project.
What good is it if you never use it?
What good is it if you don't shake it,
be careful baby now and don't you break it!
What good is it, if I can't put it on a piece of bread?
It ain't no good to nobody!
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config: 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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
--
Paola Di Maio
School of IT
www.mfu.ac.th
*********************************************
_________________________________________________________________
Message Archives:
http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config:
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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
--
---------------------------------------------------------------------
IHMC
(850)434 8903 or (650)494 3973 home
40 South Alcaniz St.
(850)202 4416 office
Pensacola
(850)202 4440 fax
FL 32502
(850)291 0667 cell
phayesAT-SIGNihmc.us
http://www.ihmc.us/users/phayes
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config: 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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (01)
|