Re: need information.

Iris,

you wrote:
> i"m looking some answers about the closed world assumption.
>  i want to know it's hisory, it's state today, it's relation to fuzzy
> logic, to ontology ,to AI ,to nonmonotonic reasoning.
> if you can't help me with that information please show me places where i
> can read that information on the web or a newsgroup in that subject

These are very good questions ...
With respect to the state of the art, you can go to http://www.google.com/
and look for "closed world assumption" or so (also combinations such
as "closed world assumption" + "fuzzy logic") and you will find
lots of relevant and vey good ranked pointers.
(Google is really good and the best I found so far).

E.g. from http://www.cs.huji.ac.il/~ai/lect11/tsld009.htm
  """Closed-world assumption (CWA)
     if you cannot prove P or ~P from KB, add ~P to KB
     Idea: if you cannot prove P, assume it is false.
     This means you assume you know everything there is
     to know about the world (e.g. the world is closed).
     This is the semantics of databases and of PROLOG."""

With respect to our works, we rather stick with classical logic
(we are not involved with fuzzy logic).

... what a nice coincidence with what I did last night ...
Instead of saying (in XSB prolog)
  :- auto_table.
  sibling(Child1, Child2):-
    childIn(Child1, Family),
    childIn(Child2, Family),
    tnot(equals(Child1, Child2)).
  equals(X,X).

one could assume the sibling property to be irreflexive and say (in N3)
  @prefix gc: <http://www.daml.org/2001/01/gedcom/gedcom#>.
  @prefix log: <http://www.w3.org/2000/10/swap/log.n3#>.
  @prefix : <#>.
   {{:child1 gc:childIn :family.
     :child2 gc:childIn :family} log:implies
  {:child1 gc:sibling :child2}}
    a log:Truth; log:forAll :child1, :child2, :family.

So in such cases one could avoid the negation (as failure) and
one also does't suffer from the *premisse ordering problem*
which can easily lead to
  ++Error: FLOUNDERING: tnot/1 used with a non-ground subgoal:

In http://www.agfa.com/w3c/euler/ we now assume properties
to be irreflexive unless otherwise declared with e.g.
  @prefix e: <http://www.agfa.com/w3c/euler#>.
  @prefix : <#>.
  :some-property a e:reflexiveProperty.

I think that one can avoid more problems with negation if
appropriate concepts are used in the ontology layer and I
think that many people on the www-rdf-logic list can give
strong evidence of this ...

--
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/

Received on Sunday, 25 March 2001 07:32:56 UTC