Copy Link
Add to Bookmark
Report
AIList Digest Volume 4 Issue 160
AIList Digest Tuesday, 1 Jul 1986 Volume 4 : Issue 160
Today's Topics:
Queries - Expert Systems for Classification & Intelligent Databases &
Constraint-Propagation Inference Engines,
Education - Special J. Instructional Science Issue,
Review - The Evidence of the Senses,
Philosophy - Creativity and Analogy
----------------------------------------------------------------------
Date: 27 Jun 1986 10:55-PDT
From: balaji@usc-cse.usc.edu
Subject: Expert systems for evolutionary classification of fish
I am sending this message for a friend. I will forward replies to her.
Thanks.
Balaji
I am writing a LISP program to classify fish and determine their
evolutionary history. This involves nesting species of fish into a
hierarchical taxonomic arrangement, on the basis of characteristics shared
among the species. My program will need to determine whether characteristics
are primitive or advanced, in order to determine how they fit into the
taxonomic hierarchy. This requires some degree of heuristic
reasoning and this part of the program will probably be constructed as a
mini expert system. The nesting process (i.e. arranging the species in a
taxonomic hierarchy) is a more or less standard procedure, but requires
flexibility to utilize heuristics in some cases.
To get a better idea of how to go about writing my program, I would
like to find out about existing programs that deal with issues similar
to mine. If anyone has any suggestions about where I might find AI
systems that might help me, or papers on such systems, I would be most
appreciative if they send me a message.
Thank you.
Noelle Sedor
------------------------------
Date: Sun, 29 Jun 86 23:27:21 CDT
From: wucs!wucec2!grs0473@seismo.CSS.GOV (Guillermo Ricardo Simari)
Subject: Intelligent Databases
If you could add only one feature to a commercial relational
DBMS in order to make it more "intelligent",
what would be your choice?
If I got enough answers I'll post a summary.
+------------------------------------------------+
| ihnp4!cuae2!ltuxa!we53!wucs!wucec2!grs0473 |
| |
| Guillermo R. Simari |
| P.O.Box 3257 |
| Saint Louis, MO 63130-0657 |
+------------------------------------------------+
------------------------------
Date: 30 Jun 86 09:27 EDT
From: Siems @ DCA-EMS
Subject: constraint propagation inference engines
jerry feinstein and david bailey of booz, allen, and hamilton
are interested in any work being done in the area of constraint
propagation that might be applied to inference engines. the specific
interest is in the use of tight constraints and a non-optimizing,
simplex-like algorithm to find a quick, "satisficing" solution in
an ordered, though not necessarily numeric, problem space. this is
a follow-up on the constraint propagation workshop held at the
expert systems conference in avignon in april of this year. any
information on current work in this area or on the use of constraint
propagation in inference engines would be greatly appreciated.
thank you.
david bailey
booz, allen, and hamilton
4330 east west highway
bethesda, md 20014
(301)951-2155
------------------------------
Date: Wed, 25 Jun 86 09:50:07 edt
From: Bob Lawler <rwl1%gte-labs.csnet@CSNET-RELAY.ARPA>
Subject: Notice of journal issue
[Forwarded from the AI-Ed Digest by Laws@SRI-AI.]
Dear Colleagues,
Today I received from Elsevier a special issue of the Journal of
Instructional Science on the theme of "AI and Education". This double-
number volume (several hundred pages in length) was prepared by
Masoud Yazdani (University of Exeter) and myself (Bob Lawler) as a
preliminary collection of articles prepared for the Second International
Conference on AI and Education held at Exeter University in September
1985. The issue is Volume 14, Nos. 3 and 4, dated May, 1986. A more
comprehensive book on the theme will be forthcoming at the end of 1986.
The contents of the special issue are as follows:
M. Yazdani and R. Lawler AI and Education: an overview
A. DiSessa Artifical Worlds and Real Experience
W. Feurzeig Algebra Slaves and Agents in a Logo-based
Mathematics Curriculum
R. Lawler and G. Lawler Computer Microworlds and Reading
H. Lieberman An Example Based Environment for Beginning
Programmers
S. Ohlsson Some Principles of Intelligent Tutoring
J. Self The Application of Machine Learning to
Student Modelling
A. Priest Solving Problems in Newtonian Mechanics
G. Drescher Genetic AI: translating Piaget to Lisp
K. Carley Knowledge Acquisition as a Social Phenomenon
If you are interested in having a copy of this journal, write to:
Elsevier Science Publishers
Science and Technology Division
P.O. Box 330
1000 AH Amsterdam
The Netherlands
The price for this double-issue of the journal is $57.25, which
includes air transport to the US and surface mail on the continent.
Bob Lawler
(LAWLER at GTE-LABS on CSNET)
(LAWLER at MIT-OZ through MIT-MC on ARPANET)
------------------------------
Date: Sat, 28 Jun 86 20:41:12 -0200
From: Eyal mozes <eyal%wisdom.bitnet@WISCVM.ARPA>
Subject: Book review: "The Evidence of the Senses"
"The Evidence of the Senses" by David Kelley
Louisiana State University Press, 1986, 262 pp., $27.50
"The Evidence of the Senses: a Realist Theory of Perception" is a
comprehensive philosophical treatment of perception, integrating
classical and recent work in philosophy and psychology. To those who
agree with its conclusions, it offers a sound, detailed framework for
psychological, biological and AI work in perception; to those who
don't, it offers an illuminating, profound and thought-provoking
alternative theory.
Dr. Kelley is formerly an assistant professor of philosophy and member
of the Cognitive Science program at Vassar College, and currently a
senior research fellow of the Ayn Rand Institute. His work is based on
the philosophy of Objectivism.
Almost all contemporary work in the theory of perception, including the
writings of philosophers, is devoted to detailed consideration of
specific issues, while taking for granted a wider context of basic
philosophical assumptions. In sharp contrast to this procedure, Dr.
Kelley makes his own basic assumptions fully explicit, defends them on
general philosophical grounds, and only then applies them to specific
issues. This makes it possible for him, when arguing against opposing
views, to argue in terms of essentials by recognizing the basic - often
hidden - assumptions on which these views and the arguments for them
rely.
A central theme of the book is the rejection of the "diaphanous model
of awareness" - the view that awareness of objects can't be mediated by
any process whose nature affects the way the objects appear; Dr.
Kelley demonstrates that this model has been accepted, explicitly or
implicitly, by almost all philosophers of perception since Kant, and it
is the root of all three common views of perception: naive realism,
which claims that our sensory apparatus is indeed diaphanous, and has
no effect on the appearance of external objects; representationalism,
which claims that we don't perceive external objects, but internal
representations which give information about these objects; and
idealism, which denies the existence of external objects.
Chapter 1 sets up the general epistemological framework for the book;
Dr. Kelley contrasts the diaphanous model with his own basic
assumption, "the primacy of existence" - the principle that
consciousness is the faculty of perceiving existence - which dispenses
with the need for making any prior assumptions about how consciousness
"should" work.
Chapters 2 through 5 apply this principle to perception. Chapter 2
deals with the relation between perception and sensation; Dr. Kelley
challenges the "sensationalist" approach - including its modern
"computational" version - which claims that perception is a process of
inference on sensations; he provides philosophical support for James
Gibson's theory of "direct perception" - which holds that external
objects are perceived directly, and that perception is a distinct form
of awareness, not composed out of sensation - and answers the major
criticisms against Gibson.
Chapter 3 treats the relation of an object to its sensory qualities.
The treatment is based on Ayn Rand's concept of "form of awareness",
which designates all perceived qualities which are relative to the
perceiver, distinguishing them from the perceived object and its
intrinsic properties; Dr. Kelley uses this concept to demonstrate the
consistency of perceptual relativity with direct realism, and
illustrates the principle in a discussion of visual illusions and in a
detailed treatment of colors; he then treats in this framework the
traditional distinction of primary vs secondary qualities.
Chapter 4 uses the principles established in previous chapters to
answer the major arguments for representationalism; this includes a
discussion of hallucinations and their relation to perception.
Chapter 5 concludes the discussion of perception by giving a full
definition - "perception is direct awareness of discriminated entities
by means of patterns of energy absorption by sense receptors" - and
discussing in detail each element in the definition and its
implications for each of the five senses.
Chapters 6 and 7 deal with perceptual knowledge, and the role of
perception as the base of conceptual knowledge. Chapter 6 discusses the
two common theories about the nature of justification: the
"foundational" theory, which holds that propositions about experiential
states are self-justifying and provide the foundation on which all
other knowledge is built as a hierarchy; and the "coherence" theory,
which holds that no single proposition can be justified outside the
context of the rest of a man's knowledge, and that the only way to
justify knowledge is by its self-consistency. Dr. Kelley identifies and
challenges the common premise implicit in both these positions - "the
propositional theory of justification", which holds that the only way
to justify a proposition is by inference from other propositions.
Chapter 7 deals with "perceptual judgments" - conceptual
identifications of perceived entities and their attributes. Dr.
Kelley's treatment of this subject is not complete, and he does not
offer a full theory; but he does indicate the direction such a theory
should take, and its implications for concept-formation. He discusses
the relation between the perceptual discrimination of an entity and the
reference to it in a perceptual judgment; the difference between
"construction" and "discovery" models of concept-formation, and their
relation to the possibility of justifying a perceptual judgment without
need for an inference from other propositions; the implications of
perceptual relativity for forming concepts of sensory qualities; and
the autonomy of perception, answering the various philosophical and
scientific arguments for the claim that perception and perceptual
judgments are affected by previous knowledge or desires.
The book is thoroughly organized, with careful attention to integration
of the various issues and to illustration of the abstract points; the
result is that, despite its highly technical content, it is very
readable. All technical terms are carefully explained, and therefore,
while reading the book will be easier for those with a previous
background in the theory of perception, such a background is not
necessary. The book contains extensive surveys of previous work and of
different views and arguments, with heavy use of references, and this
makes it an ideal starting-point for a study of the subject.
In conclusion, I strongly recommend this book to anyone seriously
interested in the theory of perception, and I think it is a must read
for any psychologist, biologist or AI researcher whose work involves
this subject.
Eyal Mozes
BITNET: eyal@wisdom
CSNET and ARPA: eyal%wisdom.bitnet@wiscvm.ARPA
UUCP: ..!ucbvax!eyal%wisdom.bitnet
------------------------------
Date: Fri, 27 Jun 86 11:23:52 edt
From: Jay Weber <jay@rochester.arpa>
Reply-to: jay@rochester.UUCP (Jay Weber)
Subject: Re: Creativity and Analogy
> Yes, I do want to understand "creativity" in terms of less slippery
>concepts, such as "analogy". We are forced to start with informal
>approaches but hope to find more formal definitions. I do not
>understand why a formal approach would satisfy very few people or
>why an informal approach would serve no useful purpose.
Consider the following view of analogy, consistent with its formal
treatment in many sources. A particular analogy, e.g. that which
exists between a battery and a reservoir, is a function that maps
from one category (set of instances) to another. Equivalently we
can view this function as a relation R between categories, in this
case we have a particular kind of "storage capability". This relation
is certainly
1) reflexive. "A battery is like a battery" (under any relation)
2) symmetric. "A battery is like a reservoir" implies
"A reservoir is like a battery" under the same relation R
3) transitive. "A battery is like a reservoir" and
"A reservoir is like a ketchup bottle" imply
"A battery is like a ketchup bottle" WHEN THE SAME
ANALOGY HOLDS BETWEEN THEM (same R).
Then any analogy R is an equivalence relation, partitioning the space
of categories. Each analogy corresponds to a node in an abstraction
hierarchy which relates all of the sub-categories, SO THE SPACE OF
ANALOGIES MAPS ONTO THE SPACE OF ABSTRACTIONS, and so under these
definitions analogy and abstraction are equivalent.
Now to the point: I recently presented this sketched proof to my peers
and they fought me whenever I tried to say "this is what analogy is"
rather than "this is what I define analogy to be" (with the latter claim
I probably should use a different term like R-analogy or XYZZY). I fact,
no one could agree to a particular formal definition of the term "analogy",
since we all have individual formal definitions by virtue of the fact that
we will answer yes or no when given a potential analogy instance, so we
are formal language acceptors with our senses as input. This is what I
mean by a "slippery" term, i.e. one that has drastically different
meanings depending on its user. This is why I say a formal definition
of analogy would satisfy very few people. Informal definitions are
useless because by defintion there is no notion of a valid inference
from the theory, we cannot make predictions with them and therefore
cannot do science with them (most "loose" defintions of things put
forward do have some formal properties, but one must be careful).
>I am sure that you do not imply that an analysis (formal or informal)
>of >anything< is futile. What is it about "creativity" that makes its
>analysis a no-win proposition?
"Creativity" is VERY slippery, perhaps only slightly less slippery than
"intelligence". Profit by Turing's example and keep your personal
definition of the slippery term in mind but define a new one, e.g.
Turing-test-intelligence instead of asking for a definition of the
word in usage.
Jay Weber
Department of Computer Science
University of Rochester
Rochester, N.Y. 14627
jay@rochester.arpa
------------------------------
End of AIList Digest
********************
