[Top] [All Lists]

Re: [ontolog-forum] IBM Watson on Jeopardy

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Mon, 14 Feb 2011 22:03:44 -0500
Message-id: <4D59ED10.8050807@xxxxxxxxxxx>
Pat,    (01)

I pushed the send button accidentally.  And I copied what I had
previously written at the end of this note.    (02)

But in any case, my main point was that the IBM Watson system
was more successful than Cyc in question answering because it
got down to addressing real questions from the very beginning.
They did not waste time defining millions of axioms.    (03)

Watson does have some ontology, but it is more of a taxonomy
of terms than anything as formal as Cyc.  The Xerox PARC NLP
group, which went to PowerSet (now Microsoft), had access
to all of Cyc, but the only thing they found useful was the
hierarchy, which they treated as a classification system
and ignored the axioms.    (04)

I believe that is the proper foundation for NLP and commonsense
reasoning:  A classification of all the terms in a hierarchy
with very few axioms.  Microtheories with detailed axioms for
specific problems should be developed from the bottom-up and
linked to the hierarchy where appropriate.    (05)

That is the experience from the IBM Watson project, from the
PowerSet project, and from our experience at VivoMind.    (06)

_________________________________________________________________    (07)

Pat,    (08)

Peter has asked us not to start new threads that break the
continuity of the discussion.  So I'm continuing the old thread.    (09)

> There is in my mind a vast difference between a " top-down, monolithic,
> detailed, universal ontology of everything" and a " perfect top-level
> ontology".  Perhaps John could supply examples of these two different things
> that have actually been proposed by someone.  I don't recall any.    (010)

People have proposed so many different things that it's impossible
to give any simple description or classification of all of them.    (011)

But people since Aristotle have been searching for a complete and
coherent ontology of everything.  Kant is the most famous one who
claimed to have a complete upper level that replaces Aristotle.    (012)

The largest modern attempt was Cyc, which started as a single monolithic
ontology in 1984, but it was reorganized as a modular collection of
microtheories in the 1990s.  The top-level hierarchy is still there,
but Lenat says that most of the detailed axioms have been pushed down
to the lower levels.    (013)

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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (014)

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