Re: Inference in daml

At 7:25 PM -0400 6/20/01, Geoff Chappell wrote:
>tes that would help Geoff and others with
>  > similar requirements.
>
>Thanks for the responses.
>
>I'm serializing inference rules now in rdf in a proprietary schema and
>thought it would be better to use daml so that the rules could potentially
>be understood by another processor. I'm getting the impression that that's
>beyond the intent of current daml language so even if I succeed there might
>not be much value in it (since other systems would be unlikely to properly
>interpret the rules properly).
>
>It was an interesting exercise for me though. I convinced myself (rightly or
>wrongly) that it is possible to express some implications in daml albeit in
>a somewhat cumbersome manner. Interestingly it seems that since you are able
>to express quantification over the object of a triple, to some degree the
>subject, but not the predicate that you can be pretty expressive if all of
>your statements are reified (and less so if they're not).

Wait - I think there is confusion here - Geoff, serializing the rules 
in DAML will indeed add power. You can then link the rules to 
ontological info, put restrictions on various fillers, and all the 
other things DAML allows.  What we don't have is a standardized model 
for expressing rules in general.  There are some kinds of inferencing 
that is built into the semantics - and that is what Ian and I have 
been discussing - but as a language for what you want to do, it 
should definitely be advantageous over raw RDF ( -- i.e. when you say 
"other systems would be unlikely to interpret the rules properly - 
you have that same problem, moreso, in raw RDF -- in DAML, you get 
some of this done for free, but you'll still need to do some yourself 
- in RDF you're doing it all)
  -JH
-- 
Prof. James Hendler		Program Manager
DARPA/ISO			703-696-2238 (phone)
3701 N. Fairfax Dr.		703-696-2201 (Fax)
Arlington, VA 22203		jhendler@darpa.mil

Received on Friday, 22 June 2001 14:30:18 UTC