On Feb 18, 2009, at 10:45 AM, John F. Sowa wrote: (01)
> Azamat, Pat C, et al.,
>
> Whenever anybody proposes a project to build a large formal
> ontology, I bring up the Cyc project, which was started in 1984. (02)
Excellent point. (03)
>
> In June 2003, DARPA sponsored a two-day review of the Cyc project
> by a panel of people working in AI, linguistics, and related fields.
> The general consensus was
>
> * Cyc is a unique and valuable resource: Since 1984, 650 person
> years and $65 million had been spent to define and axiomatize
> about 600,000 concept types.
>
> * Support for Cyc should be continued, and it should be made
> freely available for research purposes.
>
> * The Cyc researchers had acquired a great deal of experience
> about the process of developing and using a large formal
> ontology, and their successes and failures should be
> documented in a series of detailed reports. (04)
Indeed. Everyone on this forum who has an opinion about general
ontologies should read them. (05)
>
>
> In the early 1990s, the original monolithic hierarchy of the
> Cyc ontology was reorganized as a collection of "microtheories":
>
> * Microtheories are subontologies that may be inconsistent
> with one another.
>
> * As of 2003, Cyc had about 6,000 microtheories.
>
> * Cyc can create new microtheories dynamically to represent
> modalities or some agent's knowledge and belief.
>
> * But there is a need for different microtheories even at the
> upper levels of the ontology. (06)
And, critically, a given name can have different meanings in different
microtheories (so they can be called "contexts" and indeed often are.
The basic Cyc logic of microtheories was based on McCarthy's context
logic, in which facts are asserted relative to a context.). Cyc also
has its own quite complex ontology of microtheories themselves, which
are not just a set, but a complicated system with various kinds of
internal relationships and categorizations. (07)
>
>
> There were also many questions about the relationship of Cyc to
> other R & D efforts. Two of the reviewers were George Miller,
> who started the very widely used and much cheaper WordNet, and
> Chuck Fillmore, who had started the Framenet project. Following
> is the consensus about those projects:
>
> * Lexical resources such as WordNet and Framenet have different
> goals, but they are complementary to Cyc.
>
> * They have proved to be valuable for natural language projects.
>
> * The amount of investment in those projects has been modest,
> but the ROI has been considerable. Extending and linking them
> to Cyc and would require relatively modest funding.
>
> * Many related issues should be explored.
>
> One of the reviewers was Ed Feigenbaum, who had been Lenat's
> thesis adviser at Stanford. Feigenbaum was also one of the chief
> proponents and advisers for the Cyc project from the earliest
> proposals. During the discussion period, he asked a question:
>
> * Lenat had claimed that when the knowledge base reached a
> critical size, new knowledge could be added much faster.
>
> * Recently, the size of the KB has increased significantly.
>
> * Has Cyc now reached a critical mass that would support
> an exponential increase in size? (08)
I have heard Doug claim several times that Cyc is now at or near this
'turning point' where it is easier to add new knowledge or even have
the system learn for itself by reading the news feeds. I remain
unconvinced, but the sheer quantity of stuff that Cyc now "knows" is
indeed quite amazing. (09)
>
>
> Another reviewer was Fritz Lehmann, who had worked on Cyc for
> several years. He answered Feigenbaum's question: The major
> reason for the recent increase was a managerial decision.
> Lenat wanted to make the knowledge base seem more impressive
> and asked the Cyclers to add a lot of easily defined concepts. (010)
They did it by systematically stripping facts from such things as
geographical concordances, thesauri and the like. However, the fact
they were able to do even this relatively quickly is itself worthy of
note, and indicates a degree of maturity and completeness in the
underlying ontological framework. (011)
>
>
> After that review, DARPA reduced the funding for Cyc and
> required them to document their research. For copies of
> their reports, see
>
> http://www.cyc.com
>
> They also began the Cyc Foundation to make the subset OpenCyc
> freely available. See
>
> http://www.cycfoundation.org/
>
> AA> But the real fact is that the project of Standard Ontology
>> is not only a historically unique scientific and engineering
>> enterprise, but also too extensive, both in its scale of
>> knowledge, funding stakeholders and research participants.
>> Thus it will be expensive...
>
> It's only unique if you ignore Cyc, which also had a very large
> amount of knowledge, many funding stakeholders (including MCC,
> Bellcore, Eastman Kodak, pharmaceutical companies, Microsoft,
> government agencies, etc), and many of the best researchers
> in the field as employees, consultants, participants, users,
> and reviewers.
>
> PC> The estimated cost, $30M, may be low, but that is not because
>> of the size of the ontology required, but the number of
>> participants needed to be sure that the FO can translate among
>> the alternative representations that the participants may want
>> for their local use.
>
> The Cyc project was initiated in 1984 with goals that are
> a superset of anything that I have ever heard for the FO.
> Is there anything about the FO that is different from Cyc?
> If so, what and why?
>
> The groups that might fund the FO know the Cyc project very,
> very well. And they know that they never got much, if any
> return on their investment. You can't get a penny from them
> unless you can explain in detail...
>
> 1. What are you proposing that is different from Cyc?
>
> 2. What makes you think that you can be more successful
> than Cyc?
>
> 3. Why don't you start with Cyc as a foundation for the FO? (012)
Or at least OpenCyc. (013)
>
>
> We've already heard the arguments that Cyc isn't open source.
> But the cost of acquiring and adding more material to OpenCyc
> would be vastly less than $30 million. (014)
Quite. And in any case, Cycorp is now a real company, with a board of
directors, and they might be willing to come to an accommodation for a
reasonable price to make the whole of Cyc available for a suitable
long-term project, provided some protections were included for their
quite large asset. (015)
Pat (016)
>
>
> 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
> To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>
>
> (017)
------------------------------------------------------------
IHMC (850)434 8903 or (650)494 3973
40 South Alcaniz St. (850)202 4416 office
Pensacola (850)202 4440 fax
FL 32502 (850)291 0667 mobile
phayesAT-SIGNihmc.us http://www.ihmc.us/users/phayes (018)
_________________________________________________________________
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 (019)
|