John Sowa wrote: That's another way of saying "Let's redo Cyc, but not make the same mistakes" --
but without any clear idea of what mistakes, if any, Cyc had made.
What I have been saying, in an elevator pitch:
1. Cyc did a good job of defining a very large ontology.
2. I have not seen any ideas for defining a large ontology
that are significantly better than what was done for Cyc.
3. Let's look at successful applications that use some
ideas from or about ontology -- and ask (a) what made
them successful, (b) how do they differ from Cyc, and
(c) what lessons can we learn?
Cyc made a big clear mistake. It expressed its knowledge in a formal language which
was deficient on 7 simplicity-related leading edges compared with a language design
distributed at IBM in 1973. That has a big impact on inferencing because advanced
inferencing is usually applied to information about information where extraneous
complexity from poor language is compounded and the inferencing procedures are left gasping.
I would propose translating the Cyc ontology into a better language, redeveloping
the Cyc inferencing capabilities to exploit the simplicity, and seeing what improvement is made.
Ed Lowry
http://users.rcn.com/eslowry
On 8/24/2014 3:36 AM, John F Sowa wrote:
Leo and Rich,
I am most definitely *not* giving up. And I am *not* claiming
that the task is prohibitively expensive. But what I am trying
to say is that we have to do three things: (1) pause to reflect,
(2) think hard, and (3) get a some idea of what our goals should be.
Leo
What we try to do is impossible, no? And so, let's just give up, eh?
No. Cyc has devoted 30 years, over a person-millennium of work,
and well over $100 million of research money to develop the world's
largest formal ontology. And they have had many very good people
working on their project.
Any AI project that has been able to continue for 30 years deserves
a lot of credit. They have developed some useful applications. But
their ROI will not convince anyone to attempt anything similar.
Rich
John is simply emphasizing that it is an expensive undertaking
in economic terms (lots of person-hours, lots of dollars, thus
lots of people, and lots of politicality).
Not quite. I have been trying to say that we have to rethink
what *it* should be. Some people are still claiming that we need
a universal ontology of everything. That's another way of saying
"Let's redo Cyc, but not make the same mistakes" -- but without
any clear idea of what mistakes, if any, Cyc had made.
What I have been saying, in an elevator pitch:
1. Cyc did a good job of defining a very large ontology.
2. I have not seen any ideas for defining a large ontology
that are significantly better than what was done for Cyc.
3. Let's look at successful applications that use some
ideas from or about ontology -- and ask (a) what made
them successful, (b) how do they differ from Cyc, and
(c) what lessons can we learn?
John
_________________________________________________________________
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