[ont.events] Seminar on default reasoning by David Poole at Univ Toronto

gh@utcsrgv.UUCP (Graeme Hirst) (09/23/84)

 The Department of Computer Science, University of Toronto, presents:
 -------------------------------------------------------------------

      The Logical Definition of a Deduction System for Default Reasoning
				    or
	     A Machine-Oriented Proof Procedure for Default Logic

			       David Poole
		     Department of Computer Science
			 University of Waterloo


	  Tuesday 25 September 1984, 3pm, Sandford Fleming 1105
			

			      ABSTRACT

``It's the exception that proves the rule'' is an axiom of folklore that says
if a piece of knowledge is truly useful, and not just definitional, it will
have exceptions.  Previous attempts to deal with such exceptions using default
logics have resulted in systems without model theoretic semantics, and/or
without computationally acceptable proof procedures.

   This talk presents the model-theoretic semantics for a new form of default
logic. This is based on defaults being possible hypotheses in a ``scientific''
theory used to explain the results.  A machine-oriented proof procedure for the
first-order predicate calculus is outlined which combines the advantages of
goal-directed deduction systems, connection-graph proof procedures and
non-chronological backtracking. It is then shown how this can be extended to be
a proof procedure for default logic in a natural way.

   Why this is better than other proposals; how the problems that give rise
to non-normal defaults are overcome; and how this relates to diagnosis expert
systems are also discussed.
-- 
Graeme Hirst

henry@utzoo.UUCP (Henry Spencer) (09/24/84)

>``It's the exception that proves the rule'' is an axiom of folklore that says
>if a piece of knowledge is truly useful, and not just definitional, it will
>have exceptions.

Actually, whoever wrote this abstract (Dr. Poole?) should be embarrassed,
because he's got it wrong.  The saying is so old that the meaning of the
words has changed.  "Prove" is being used in its obsolete meaning of
"to test".  Translated to modern English, the saying reads:  "Extreme
cases are the best tests of a rule".  Tsk, tsk.
-- 
				Henry Spencer @ U of Toronto Zoology
				{allegra,ihnp4,linus,decvax}!utzoo!henry