Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!ucbvax!NUSVM.BITNET!ISSSSM
From: ISSSSM@NUSVM.BITNET (Stephen Smoliar)
Newsgroups: comp.ai
Subject: RE: UNIFIED MODEL FOR KNOWLEDGE REPRESENTATION? (IMPOSSIBLE
Message-ID: <9106040003.AA12879@lilac.berkeley.edu>
Date: 3 Jun 91 23:59:46 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Lines: 116
X-Unparsable-Date: Tue, 04 Jun 91 08:00:10 SST

In article <5255@syma.sussex.ac.uk> aarons@syma.sussex.ac.uk (Aaron Sloman)
writes:
>
>Some of the people who have worked on first order logic have thought
>it could serve as the universal (= unified?) notation for
>representing knowledge. I don't know if this is what you mean by a
>model.
>
>However, it seems pretty obvious from the history of science and
>culture that different formalisms are useful for different purposes,
>including, logic, algebra, pictures, maps, tables, flow-charts,
>musical notation, 3-D models, natural languages, etc.
>
>One reason for this is that different kinds of notations have
>different kinds of variability, which limits their expressive power
>in different ways. This can sometimes be important when exploring
>large search spaces. If the structure of the notation is such that
>it won't let you express things that would only be rejected anyway,
>it can have great heuristic power.
>
>If anyone claims to have a unified model, ask him/her if it will
>serve for the purposes of representing knowledge about the contents
>of the current optic array in a robot's visual system, and for
>transforming that knowledge in the process of discovering what's out
>there (e.g. discovering binocular disparities for stereo vision) or
>for fine control of posture and actions, etc.
>
In article <1991Jun3.110003.13773@msuinfo.cl.msu.edu>
sticklen@pleiades.cps.msu.edu (Jon Sticklen) then runs
with this ball, writing:
>
>The point of what in knowledge-based systems is being called "task specific
>architectures" (TSAs) is to suggest that problem solving is best analyzed and
>"mimiced/implemented" by using primitives that are used in the domain of the
>problem solving. For example, in a diagnostic domain, it is natural to both
>analyze diagnostic problem solving, and to build computer versions of
>diagnostic problem solving, by using the concept of "diagnostic hypothesis".
>Specializations of the TSA concept like Chandrasekaran's generic tasks (GTs)
>take one additional step of saying that there exists a finite set of problem
>solving strategies that are generally useful. These GTs are defined by giving
>both a knowledge representation template, and a control strategy. For example,
>there is a GT for classification problem solving, one for simple ("routine")
>design, one for function-based model level reasoning...
>
>At the same time as researchers seek to extend the capability of TSAs like
>those in the generic tasks, grand architectures for problem solving are being
>developed to handle any problem solving situation - SOAR may be the easiest
>example. SOAR generates problem spaces appropriate for a given problem.
>Although not a developed capability yet, the SOAR architecture may one day be
>able to generate very tailored problem spaces after analyzing a given problem
>situation - perhaps problem spaces not unlike GTs.
>
>If that happens, then it will be a unification. But a unification along the
>lines that Arron pointed to, not a unification at the base level of saying
>"one
>representation fits all."
>
I think this approach may be missing the point Aaron was trying to make.
Generic tasks and problem spaces are, once the dust of surface appearance
is swept aside, just as much mathematical objects as are the constructs of
first-order logic.  The difficulty lies in attempting the act of reductionism
itself, rather than in the particular construct which is the target of your
reduction.

Aaron offered up a nice list of different formalisms which have been engaged
for different purposes of reasoning.  However, the more we begin to recognize
the importance of viewing reasoning as SITUATED, the more difficult it becomes
to carve off "the reasoning itself" as an object of study.  This opens the door
to the intimidating prospect that we need to be concerned with more general
issues of BEHAVIOR, which means that the formalisms in Aaron's list are only
scratching the surface.

Consider this problem from another point of view.  Regardless of whether or not
we think it is feasible to build machines which satisfy "behavioral criteria
for intelligence" (whatever those criteria may be), let us simply worry about
whether or not we can DESCRIBE intelligent behavior.  Let me say right off that
I shall be the first to raise a skeptical eyebrow at any claim that this
problem has been solved in any practical way.  The fact is that psychology
is still, to a great extent, floundering around simply trying to DESCRIBE
many of the phenomena it wishes, ultimately, to explain.  For many limited
domains of scientific reasoning, we can at least fall back on foundations
of formalisms such as those Aaron has enumerated;  but when we try to take
on behavior (even when intelligence isn't directly involved), we might as
well be in a raft in the middle of the North Atlantic Ocean.

I would like to pose a possibly radical explanation for why we find ourselves
so lost here, and that is that THERE IS NO SUCH THING AS OBJECTIVE DESCRIPTION.
Description is highly subjective to the person doing the describing.  If anyone
else wishes to draw upon that description, he is obliged to enter into a
relatively sophisticated process of negotiation which is known as communicating
in a natural language.  (Anyone who is interested in seeing blatant examples of
how subjectivity lurks in seemingly objective accounts should take a look at
"Apes R Not Us," a review of primate studies by Lord Zuckerman which appeared
in the May 30 issue of THE NEW YORK REVIEW.  Also relevant is Brian Smith's
response to Lenat and Feigenbaum, "The Owl and the Electric Encyclopedia,"
published in Volume 47 of ARTIFICIAL INTELLIGENCE.)

In the face of such a daunting proposition, it may be that the quest for a
"unified representation" is not so much misguided because of the "unified"
attribute as because of the very goal of representation itself.  We are now
entering a period of skepticism regarding the issue of representation.  Such
skepticism is encouraged by the work of Smith, Rodney Brooks, and others far
too numerous to mention.  If any of these researchers can demonstrate results
which scale up from small experiments to practical problems of getting on in
the world, our current obsession with representation may ultimately be
dismissed as a distracting side-track.

===============================================================================

Stephen W. Smoliar
Institute of Systems Science
National University of Singapore
Heng Mui Keng Terrace, Kent Ridge
SINGAPORE  0511

BITNET:  ISSSSM@NUSVM