Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!tdatirv!sarima
From: sarima@tdatirv.UUCP (Stanley Friesen)
Newsgroups: comp.ai
Subject: Re: What AI is Exactly - Another Update.
Keywords: intelligence
Message-ID: <148@tdatirv.UUCP>
Date: 19 Sep 90 19:52:49 GMT
References: <3734@gara.une.oz.au>
Reply-To: sarima@tdatirv.UUCP (Stanley Friesen)
Organization: Teradata Corp., Irvine
Lines: 102

In article <3734@gara.une.oz.au> pnettlet@gara.une.oz.au (Philip Nettleton) writes:
>[it is] time to repost the definition of an "Intelligent System" that we have
>developed so far. Pinning this debate back to its origins, we would
>be interested in hearing from anyone with a CONSTRUCTIVE critism of
>any part of the definition or any additions they feel are necessary.
...
>Let us produce a slightly more refined "general requirements" for the
>behaviour of an "intelligent system".

O.K here goes, my general comments on the definition.

In general I think it is very good.  It seems to capture in a fairly clear
and concise way most of my intuitive definition of intelligence.

Now for some detail comments:

>			DEFINITION:
 
>a)	The system MUST be able to learn. This implies that the
>	system MUST have a memory for learning to be maintained.

>	ii)	It SHOULD be able to learn by instruction, but this
>		is not necessary. At the very least the system MUST
>		have preprogrammed instincts. This is a boot strap
>		for the developing intelligence.  Without a starting
>		point, the system cannot progress.

I rather suspect that instruction will turn out to be a special case of
learning from observation.  Or it may be a composite of all three of the
other modes of learning.  I certainly doubt it is actually a distinct mode
in its own right.  However, mentioning it here is probably appropriate, since
this is neither certain, nor entirely obvious.

I think some sort of definition of 'instinct' might be in order below.  It
should probably be fairly general, so as to allow for a wide variety of
implementations. [For instance, a chess program with prewired rules and
basic moves (including opening books), could be said to have an instinctive
knowledge of chess - if it then could expand it repertoire, and improve
its game over time, I would say it was learning]

>b)	The system MUST be autonomous.  ...
>	This can be disected as:
 
>	i)	The system MUST be able to effect its environment
>		based on its own independent conclusions.

Perhaps 'internal processes' might be better than 'independent conclusions'.
As written this assumes a particular class of models of decision making.  I
am not sure that this may not be too restrictive.

>
>	ii)	The system MUST be its own master and therefore
>		doesn't require operator intervention.

Maybe a little more precision about what it means to 'be its own master'.
This could be construed so as to rule out any form of subordinate relationships,
such as employees, servents, slaves &c.   To say that human slaves are/were
unintelligent is certainly not what you intended.  [I suspect I agree with
what you really mean, but it is not really clear here]

>	iii)	The system MUST be motivated. It must have needs and
>		requirements that can to be satisfied by its own
>		actions.

This is one that may not really be necessary.  It could either be an emergent
result of all the others, or it could be irrelevant to intelligence.
I am sonewhat uncertain here - does anyone have any other comments.

>c)	The system MUST be able to reason. That is to say, it must
>	use some form of deductive reasoning, based on known facts
>	and capable of producing insights (deductions) which later
>	become known facts.

Limiting reasoning to deductive reasoning is almost certainly too restrictive.
Most existing intelligent animals, and humans in particular, use analogical
reasoning much of the time (based on similarities and patterns rather than
principles and derivations). Almost any sort of extrapolation should be
included.  Also, internal 'manipulation' of mental models of reality to
estimate the effect of various actions is one of humanities most powerful
forms of reasoning.  We call it "imagination", and "mental rehersal" and
"planning" and many other things.

I do agree that reasoning is a critical component of intelligence, but the
defintion of resoning needs to be general enough to cover most types used
by humans in day-to-day activities.

>d)	The system MUST be able to develop self awareness. This is
>	related to autonomy, reasoning and learning, but also
>	embodies the need for external senses. Without external
>	senses there is no way of appreciating the difference between
>	"me" and "outside of me". Sensationations of pain and
>	pleasure can provide motivation.

I have mentioned this before, but I believe self awareness comes from self-
monitoring.  That is from 'sensory' input about internal state as well as
external environment.  In humans, and probably most animals, this includes
a kinesthetic sense, pain, hunger (and other bodily need sensitivities).
Self-awareness would then develop from the observation that the internal
senses follow different rules than the external ones.

-----------------------------
uunet!tdatirv!sarima				(Stanley Friesen)