Re: Concrete OpenCyc Concerns (was: RE: SUO: ELP's summary of MRW's standards experience)
Erik,
At 09:37 AM 6/24/2003 -0400, Eric Peterson wrote:
>Thanks for your reply, Adam.
>
>Can you point us to some small number of OpenCyc axioms that are
>incomplete in the absence full-Cyc axioms and explain precisely what is
>lacking in the SUO context?
Sure, take a look at almost any term and read its definition. You'll find
that there are not formal axioms that contain the definitional content in
English. For example
(#$comment #$touches "(#$touches THIS THAT) means that THIS and THAT are in
contact, either directly or indirectly. Two spatial things are directly in
contact just in case there is nothing between them; ...")
You'd like to have an axiom that says just that:
(implies
(touches ?THIS ?THAT)
(not
(exists (?OTHER)
(between ?THIS ?THAT ?OTHER))))
Some of the semantic content needed for the ontology is there in the base
axiom set that includes mathematical things like transitivity, and some
semantics is embodies in the type structure, but a lot is missing. Cyc
itself is an important achievement. It includes these axioms, but the
current OpenCyc release lacks this useful content. My hope is that one day
the rest of the content will also be released.
Adam
>For my previously stated thoughts on this subject:
>http://grouper.ieee.org/groups/suo/email/msg09587.html
>
>
>Cheers!
>
>-Eric
>
>
>
> > -----Original Message-----
> > From: Adam Pease [mailto:adampease@earthlink.net]
> > Sent: Monday, June 23, 2003 7:00 PM
> > To: standard-upper-ontology@ieee.org
> > Subject: RE: SUO: ELP's summary of MRW's standards experience
> >
> >
> > Eric,
> > One of my concerns about OpenCyc is that it lacks the axioms
>present in
> > the proprietary version. If we are to use an SUO for reasoning, we
>need
> > the rules that define the associated concepts. There are other
>concerns
> > that folks have voiced, but that's a clear and concrete one to start
>with.
> >
> > Adam
> >
> > At 10:24 AM 6/23/2003 -0400, Eric Peterson wrote:
> >
> > ><snip>
> > >.
> > > >
> > > > >People can learn from their own mistake, although not always. It
>is
> > > > harder
> > > > to
> > > > >learn from the mistakes of the other. Especially when you don't
>know
> > >what
> > > > they
> > > >
> > > > >are.
> > > >
> > > > Right. That's why I promise to itemize, analyze, and
> > > > criticize their mistakes in my forthcoming report.
> > > >
> > > > John
> > >
> > >[ELP] I went to a AAAI tutorial on knowledge sharing several years
>ago
> > >where the really big name presenter showed a bunch of slides on
> > >relational query optimization - which was one of his big specialties.
> > >
> > >My point here is that we all are pressed with the temptation to show
> > >existing less relevant stuff rather than taking the time and care to
>be
> > >on point.
> > >
> > >I just want to make the point that until full Cyc is made public, it
>and
> > >criticisms about it are only of attenuated relevance to any
>discussions
> > >of this work group.
> > >
> > >I want to make sure that we are properly focused on the fact that
> > >OpenCyc is the only directly relevant Cycorp artifact on the table
>for
> > >discussion.
> > >
> > >The preceding admonition may not apply to you, but I respectfully ask
> > >you to - rather than citing papers critical of Cyc - cite some small
> > >number of axiomatic criticisms within those papers that clearly and
> > >demonstrably apply to OpenCyc. Three axioms per source would be a
>great
> > >start and would give us some grounded confidence that you have the
>right
> > >to cite these papers in the SUO context.
> > >
> > >I don't think waiting 'till later is at all acceptable when you are
> > >using your non-specific Cyc criticisms to sway SUO policy in the mean
> > >time.
> > >
> > >I could have easily missed some, but the only concrete OpenCyc
> > >criticisms that I have seen surface on this group, I had to press
>for.
> > >And they turned out not to be valid.
> > >
> > >Thanks in advance for your focused attention to establishing the
> > >relevance of you Cyc criticisms.
> > >
> > >-Eric