COMMENT ⊗   VALID 00006 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	ockham[s88,jmc]		Notes April 15 Ockham's Razor lecture
C00004 00003	Everyone has common sense, but no-one knows how to program a computer
C00006 00004	slides
C00007 00005	Everything that happens has a cause
C00008 00006	outline
C00009 ENDMK
C⊗;
ockham[s88,jmc]		Notes April 15 Ockham's Razor lecture
Occam's razor for computer programs

William of Ockham in the 14th century enunciated
the philosophical principle ``Do not multiply entities
beyond necessity'', and since then philosophers have studied
how to apply it.  In the 1970s it became clear that intelligent
computer programs also need the principle, but for them it needs
to be made precise.  Since the late 1970s various forms of
``nonmonotonic reasoning'' have been formalized in mathematical
logic and used in computer programs.  

The talk will concern this interaction between engineering and
philosophy, but the emphasis will be on the engineering.


Everyone has common sense, but no-one knows how to program a computer
to have common sense.  That's not surprising; fish don't know how to
design submarines.

questions
Common sense database

	1. generality, sufficiency, qualification problem

Ripostes:
	Do you think physicists should give up on turbulence?

Examples:
	birds
	Nixon?
	Yale with causes
	attempting to bribe
	but
	the cause of the extinctions
	missionaries and cannibals, rule of conjecture and
rules of inference

relation to probability theory

What is the probability that there are no more relevant entities?
The theory is somewhat orthogonal to issues about Baysianism or
subjective probabilities.

In some sense we rehabilitate the obsolete notion that everything
has a cause.  Ask Suppes for a reference - perhaps to Russell.
slides

Turing started serious AI efforts
ai will win
ai is hard
Occam is ubiquitous, need lots of examples
formalized nonmonotonic is new
Occam's razor isn't merely a philosophical principle to be applied
when proposing theories.  It's used all the time in every day life.

missionaries and cannibals
Everything that happens has a cause
outline

AI is hard but has had some success.
	Common sense is hard
	Common sense is important
	What is common sense?
We have common sense but don't understand how it works.
For example, common sense physics has a different information
situation than physicists' or engineers' physics.
Russell's mistake.
slide on the Nemesis theory