Copy Link
Add to Bookmark
Report
AIList Digest Volume 6 Issue 053
AIList Digest Sunday, 20 Mar 1988 Volume 6 : Issue 53
Today's Topics:
Philosophy - Implications of the Chinese Room,
Application - References on VLSI Testibility Checking &
Agricultural Uses of AI,
Logic - Funny Logics and AI,
Course - Graduate Study in AI (Washington University),
Genetic Algorithms - Definition and Schools of Research
----------------------------------------------------------------------
Date: 14 Mar 88 11:03 PST
From: hayes.pa@Xerox.COM
Subject: Re: AIList V6 #51 ..implications of the Chinese Room
Adrian G C Redgers writes:
>a) I thought Searle's point was that humans might not "understand" >Chinese (or
English) and are simply manipulating symbols which >model the world. The
'Chinese room' is then a brain. ... Or was >Searle pointing out that the room is
unsatisfactory for those very >reasons?
Why not try reading Searle? He couldnt be clearer or more entertaining ( or
wronger, but thats another story ). He isnt claiming that brains arent
machines, or that humans dont understand Chinese. His point is that a
programmed computer cant understand anything, even if it behaves impeccably,
passing the Turing test all over the place. Reason: well, the program cant
because its just a pile of symbols, and the unprogrammed hardware ( =the man in
the room ) certainly doesnt know anything, being just dumb electronics, and
that/s all there is in a programmed computer, QED. A brain, now, is different,
because of course brains understand things: and the conclusion obviously is that
whatever sort of machine the brain is, it isnt a programmed computer. So
`strong AI' is wrong, Turing test and all. Weak AI, on the other hand, just
claims that it is simulating intelligence on electronics, which is fine ( says
Searle ) - probably a scientific mistake, he guesses, but not a philosophical
mistake, and immune from the Chinese room argument.
Pat Hayes
------------------------------
Date: Tue 15 Mar 88 13:05:24-CST
From: Charles Petrie <AI.PETRIE@MCC.COM>
Subject: Reply to Gabriel - VLSI Testibility Checking
The MCC CAD program has a knowledge-based testibility checker -
contact Thomas@MCC.com
The NCR Design Advisor includes testibility checking -
contact AI.Steele@MCC.com
also see Robin Steele's article in the proceedings of the 24th
Design Automation Conference (ACM/IEEE)
Both are commercial applications. NCR charges a minimum of $4,000
for users to come in and run their designs through the Design Advisor
and interact with it.
------------------------------
Date: 16 Mar 88 14:07:06 est
From: Mark Shirley <mhs@ht.ai.mit.edu>
Subject: applications of AI techniques to Testability in VLSI design?
Does anyone out there have any references or information on
applications of AI techniques to Testability in VLSI design?
Thanks in advance,
Gabriel.
AI and Design for Testability:
@InProceedings(shirley87,
Key="shirley87",
Author="Shirley, M., P. Wu, R. Davis, G. Robinson",
Title="A Synergistic Combination of Test Generation and Design for
Testability",
Organization="The Computer Society of the IEEE",
BookTitle="International Testing Conference 1987 Proceedings",
Year="1987",
Pages="701-711")
In the last couple of years, the testing conference has had AI sessions
\bibitem{abadir85}
Magdy~S. Abadir and Melvin~A. Breuer.
\newblock A Knowledge-Based System for Designing Testable VLSI Chips.
\newblock {\it IEEE Design \& Test of Computers}, 56--68, August 1985.
AI and Test Generation:
@InProceedings(Shirley86,
key="Shirley86",
Author="Shirley, M.",
Title="Generating Tests by Exploiting Designed Behavior",
Organization=AAAI,
BookTitle="Proceedings of the Fifth National Conference on
Artificial Intelligence (AAAI-86)",
Year=1986,
Month=August,
Pages="884-890")
@InProceedings(Singh86,
key="Singh86",
Author="Singh, N.",
Title="Saturn: An Automatic Test Generation System for
Digital Circuits",
Organization=AAAI,
BookTitle="Proceedings of the Fifth National Conference on
Artificial Intelligence (AAAI-86)",
Year=1986,
Month=August,
Pages="778-783")
Related DFT and Test Generation work:
\bibitem{horstmann}
Paul~W. Horstmann.
\newblock Design for Testability using Logic Programming.
\newblock In {\it Proceedings of 1983 International Test Conference},
pages~706--713, 1983.
@PhDThesis(Lai81,
Key="Lai",
Author="Lai, Kwok-Woon",
FullAuthor="Larry Kwok-Woon Lai",
Title="Functional Testing of Digital Systems",
School=CMU,
Number="CMU-CS-148",
Month="December",
Year="1981")
\bibitem{williams73}
M.~J.~Y. Williams et al.
\newblock Enhancing testability of large-scale integrated circuits via test
points and additional logic.
\newblock {\it IEEE trans on Computers}, C-22(1):46--60, January 1973.
------------------------------
Date: 18 Mar 88 08:34:47 GMT
From: munnari!metro.ucc.su.oz.au!daemon@uunet.UU.NET
Subject: Re: Agricultural Uses of AI
Siratac is an expert system for the advice of pest management for
cotton farmers. it is being developed by the CSIRO Division of
Information Technology and Division of Plant Industry in Australia.
Several references are available in conference proceedings both
computer/ai and cotton related. However, a talk is being presented to
the AAAI Workshop Series in the US this year specifically to do with
Siratac. If you want any copies contact me at the following adresses
ARPA: jansen%ditsyda.oz@seismo.css.gov
CSNET: jansen@ditsyda.oz
UUCP: {enea,hplabs,mcvax,prlb2,seismo,ubcvision,ukc}!munnari!ditsyda.oz!jansen
AUSPAC: jansen@au.csiro.ditsyda
------------------------------
Date: Tue, 15 Mar 88 13:39:55 GMT
From: Flash Sheridan <flash%ee.qmc.ac.uk@NSS.Cs.Ucl.AC.UK>
Reply-to: flash <@NSS.Cs.Ucl.AC.UK,@cs.qmc.ac.uk:flash@ee.qmc.AC.UK>
Subject: Re: Funny Logics and AI: references
You might also look at _Non-Standard Logics for Automated Reasoning_,
Academic Press, 1988, ed. E.H.Mamdani et al. It's a bunch of polemics
and discussions and expositions of lots of different approaches, of varying
quality.
------------------------------
Date: 18 Mar 88 21:22:39 GMT
From: wucs1!wucs2!posdamer@uunet.uu.net (Jeff Posdamer)
Subject: Graduate study in AI offering (Washington University)
GRADUATE CERTIFICATE IN ARTIFICIAL INTELLIGENCE
Intensive Program
Summer Session - May 30-August, 1988
This Graduate Certificate in Artificial Intelligence offers a
strong foundation in the theory, techniques and methods of AI.
Graduates will have a grounding in knowledge engineering, AI pro-
gramming methods and languages, knowledge acquisition and
representation, and application development tools and methods.
Ideal for engineers, scientists and MIS professionals, this pro-
gram will give the students an intensive preparation in the fun-
damentals of artificial intelligence and expert systems with em-
phasis on the practices required to design and construct applica-
tions. The program has over 100 graduates and is currently in
its fourth offering. Graduates will earn credit applicable to a
graduate degree and will be awarded a graduate certificate upon
satisfactory completion. the program emphasises laboratory work.
A significant part of the program is the students' projects. Each
student should enter the program with a proposed AI project. The
individual student performs knowledge acquisition, luation, en-
coding and analysis. The staff will select projects to be imple-
mented as prototypes by teams of three or four students. Final
project reports are presented in seminar and written form.
For further information contact:
Center for Intelligent Computing Systems
Campus Box 1141
Washington University
Saint Louis, MO 63130
(314)-889-6766
tonya@syr.wustl.edu
--
Jeff Posdamer, Washington University, St. Louis, MO, (314) 889-6147
posdamer@syr.wustl.edu
------------------------------
Date: Wed, 16 Mar 88 18:16:17 PST
From: rik%cs@ucsd.edu (Rik Belew)
Subject: A four-bit definition of Genetic Algorithms
My two-bit definition of "genetic algorithms has,
like all over-simplifications, drawn criticisms
for being --- you guessed it --- over-simplified!
So here's a second pass:
First and foremost, I want to "make one thing
perfectly clear": When I defined GA(1) in terms
of "... John Holland and his students" I most
certainly did not intend that only those that
had journeyed to Ann Arbor and been annointed
by the man himself were fully certified to open
GA franchises! Defining a scientific approach
in terms of the personalities involved is not
adequate, for GA or any other work. I was
attempting to distinguish a particular approach
from the broader set of techniques that I called
GA(2). In my experience, John is the "root"
of all such work and much of it has been
done by "direct" students of his at Michigan
and "indirect" students of these students. I
also know, however, that others --- notably Dave
Ackley and Larry Rendell --- have worked on
these methods without direct contact with this
lineage. But I very much consider them "students
of Holland," in that they are aware of and have
benefited from John's work. (Again, I mean that
as a compliment, not because I have been charged
with GA Clique membership validation.) I see
absolutely no benefit and potentially great harm
in drawing lines between closely related bodies
of research.
So let's move on to more meaningful attempts to
define the GA. My two-bit definition focused on
the cross-over operator: GA(1) depended on it,
and GA(2) generally relyed on the weaker (less
intelligent) mutation operator. This lead Dave
Ackley to feel that:
... membership in GA(1) is restricted
to a small and somewhat quirky "DNA-ish"
subset of all possible combination rules
[ackley@flash.bellcore.com, 3 Mar 88]
I take Dave to mean that the algorithm
presented by Holland (let's say the R class of
algorithms described in his ANAS Chapter 6, to
be specific) sacrifices some performance in order
to remain more biologically plausible. But I'm
with Dave on this one: Personally, I'm also more
intereted in the algorithms than their relation
to real biological mechanisms. (Let the record
show, however, that there are GA practioners
who do try to take biological plausibility more
seriously, e.g., Grosso and Westerdale.)
So the next possibility is to refer to
properties of the GA we find desirable. For
Dave, I think the key property of the GA is its
"implicit parallelism": the ability to search a
huge space implicitly by explicitly manipulating
a very small set of structures. Jon Richardson
comes closer to the definition I had in mind
with his emphasis on Holland's "building blocks"
notion:
The proper distinction I think is
whether or not the recombination operator
in question supports the building block
hypothesis. "Mutation-like operators"
do not do this. Any kind of weird
recombination which can be shown to
propagate and construct building blocks,
I would call a Genetic Algorithm. If
the operator does nothing with building
blocks, I would consider it apocryphal.
It may be valuable but apocryphal
nonetheless and shouldn't be called a GA.
[richards@UTKCS2.CS.UTK.EDU, 4 Mar 88]
While I would echo the value of both these
desiderata, I don't find them technically tight
enough to be useful. So I suggest that we follow
Holland's suggestion (in his talk at ICGA85)
and reserve the term "GA" for those algorithms
for which we can prove the "schemata theorem"
(ANAS, Theorem 6.2.3). I believe this theorem
is still the best understanding we have of how
the GA gives rise to the properties of implicit
parallelism, building blocks, and why. Of
course, there are problems with this definition
as well. In particular, it is so tightly
wed to the string representation and cross-over
operator that it is very difficult to imagine any
algorithm very different from the (traditional)
GA that would satisfy the theorem. But that's
exactly where I think the work needs to be done.
Finally, I want to say why I (as Dave Ackley
says) "took the trouble to exclude" Ackely's SIGH
system from my definition of GA(1). My answer
is simply that I view SIGH as a hybrid. It
has borrowed techniques from a number of areas,
the GA, connectionism, simulated annealing to
name three. There is absolutely nothing wrong
with doing this, and as Dave's thesis showed and
Larry Eshelman's note confirmed [Larry.Eshelman-
@F.GP.CS.CMU.EDU, 11 Mar 1988] there are at least
some problems in which SIGH does much better
that the traditional GA. My only problem with
SIGH is that I can't do the apportionment of
credit problem: when it works, I don't know
exactly which technique is responsible, and when
it doesn't work I don't know who to blame.
I too think about connectionist algorithms and
simulated annealing along with Holland's GA and
bucket brigade, and see all of them as members
of a class of algorithms I want to understand
better. But I find it necessary to isolate the
properties of each before trying to combine them.
In short, I think Dave and I agree to a great
extent (on the problem, and on what portions of
the solution might be), and disagree only in our
respective approaches to putting it all together.
------------------------------
End of AIList Digest
********************
