On Wed, February 20, 2013 01:44, Simon Spero wrote:
> On Wed, Feb 20, 2013 at 12:59 AM, doug foxvog <doug@xxxxxxxxxx> wrote: (01)
>> The requirement is not to express it in *FOL*. (02)
> From what I have heard, the better answers to the some of the test
> questions involved quantification over predicates (03)
I took took the test before working at Cycorp for 7 years and becoming
an "ontological architect" there. No test questions that i received,
nor ones that were given to other prospective employees (as far as
i know) required quantification over predicates. (04)
> (which is not
> traditionally GOFOL, even though Cyc and CL have a first-order semantics). (05)
Cyc is not restricted to first-order semantics. Although, of course, most
statements in Cyc are first-order statements. (06)
Cyc can quantify over predicates. Cyc also defines various properties of
predicates by using the predicates as arguments to the definitional
predicates. (07)
> Tying back in to the original subject, this kind of expressivity allows
> for compact expression of common patterns,
> but avoids many of the pitfalls in
> true higher order semantics. (08)
Yes. Many of the higher-order forms in Cyc have explicit code support. (09)
As a result of analyzing thousands of rules written in Cyc, many
regular rule patterns were discovered. We took advantage of this
by defining "rule macro predicates", being higher-arity (normally)
predicates with an argument place for each variable in the rule,
converting the rules to simple statements using the rule macro
predicate, and providing code support to speed up reasoning using
the rule. (010)
> Also on a related note, 4D and 3D+1 are just different views of the
> same rabbit. (011)
Agreed. (012)
> ...
> What matters is that the representation presented to the users of the
> system is matches their intuitive understandings, and that the
> representation used for reasoning uses is optimal for the computations
> performed. (013)
Agreed. This is a user-interface issue. (014)
-- doug f (015)
_________________________________________________________________
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 (016)
|