Copy Link
Add to Bookmark
Report
AIList Digest Volume 5 Issue 268
AIList Digest Friday, 13 Nov 1987 Volume 5 : Issue 268
Today's Topics:
Review - Spang Robinson V3 N10,
Bibliography - Leff File bm846,
Comments - Success of AI & Gilding the Lemon & FORTRAN
----------------------------------------------------------------------
Date: Mon, 9 Nov 1987 02:29 CST
From: Leff (Southern Methodist University)
<E1AR0002%SMUVM1.BITNET@wiscvm.wisc.edu>
Subject: Review - Spang Robinson V3 N10
Summary of the Spang Robinson Report on Artificial Intelligence
October 1987, Volume 3, No. 10
The lead story is on Financial Expert Systems:
A survey of insurance and companies show that 21 per cent are using
expert systems with 20 per cent having no activity and the others in
various stages of development or research. For banks, the figures are
12 and 47 per cent respectively. The article gives information on
management attitudes, uses and comparisons of activity in property and
casualty and life insurance, use of PC's, mainframes and lisp machines and
type of language.
(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^(^
Review of video tape classes on expert systems, "AI Masters"
by Addison-Wesley. This set has courses given by Patrick H. Winston,
Randall Davis and J. Ross Quinlan. The training aid has work books and
a simple PC expert tools. The workbooks have checklists to be used
in tool and application selection and test. The training system maligns,
perhaps due to datedness, PC-based expert systems and induction tools.
The three courses run for $2500-$3500 apiece with additional workbooks
for $10.00 a piece.
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
Programs in Motion's Fusion allows user to put in examples and
generate production rules. The system can accept an example matrix of
32 factors and 32 resultsants and up to 255 different examples to
generate rules. (There can be more than 255 cases if some of the cases
are redundant.)
The system does allow chaining of the decision rules. Fusion can
generate C, Pascal and production code and read in dBase files.
(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_(_
shorts:
Symantec Corporation has merged with THINK technologies.
Digitalk has released a new version of Smalltalk/V with high resolution
object oriented programming for IBM PS-2/25 and 30 computers.
Cognitive Systems, Inc. has developed a system to read messages and route
them to the appropriate people in a bank.
Teknowledge has been awarded a $1.2 million contract for work on Pilot's
Associate.
U. S. Army is purchasing ART plus various services from Inference Corporation
(more than $3 million worth)
Palladian Software has sold its Operations Advisor to Blue Cross and Blue
Shield.
Odetics got a contract to apply AI to residual heat removal in
nuclear power plants.
Gold Hill Computer has signed a distriubtion agreement with
Computer Engineering and Consulting of Japan.
System Research and Development Co. of Tokyo
has developed a new expert system building
tool called ESPARON.
_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)_)
Discussion of the Gigamos vs. Gensym dispute.
Gigamos and Gensym are both headed by former leaders of LMI who sold
all assets to Gigamos. Gigamos charges Gensym with using trade secrets
and confidential information to develop a new expert system for
real time applications (G2) in competitition with Gigamos. Gigamos
charges Gensym founders
with "planning to resign from LMI and to use LMI proprietary
information in the new GENSYM business venture." They also accuse Gensym
of causing other LMI resignations helping defeating LMI financing.
Gigamos is asking for a copy of the software and source code to be
deposited.
------------------------------
Date: Thu, 12 Nov 1987 02:53 CST
From: Leff (Southern Methodist University)
<E1AR0002%SMUVM1.BITNET@wiscvm.wisc.edu>
Subject: Bibliography - Leff File bm846
Defs for a62C
D MAG144 IEEE Transactions on Systems, Man, and Cybernetics\
%V 17\
%N 3\
%D MAY-JUN 1987
D MAG145 International Journal of Man-Machine Studies\
%V 26\
%N 2\
%D FEB 1987
D MAG149 Information Processing and Management\
%V 233\
%N 4\
%D 1987
D MAG150 Computer Vision, Graphics, and Image Processing\
%V 39\
%N 3\
%D SEP 1987
D MAG151 International Journal of Man-Machine Studies\
%V 26\
%N 3\
%D MAR 1987
D MAG152 Image and Vision Computing\
%V 5\
%N 3\
%D AUG 1987
D MAG153 Computers and Industrial Engineering\
%V 13\
%N 1-4\
%D 1987
D MAG154 Fuzzy Sets and Systems\
%V 23\
%N 3\
%D SEP 1987
D MAG155 International Journal of Man-Machine Studies\
%V 26\
%N 4\
%D APR 1987
D MAG156 Computer Vision, Graphics, and Image Processing\
%V 40\
%N 1\
%D OCT 1987
D BOOK85 Image Pattern Recognition: Algorithm Implementations,\
Techniques, and Technologies\
%S Proceedings of the Society of Photo-Optical Instrumentation Engineers\
%V 755\
%E F. J. Corbett\
%I SPIE - International Society Optimal Engieering (Bellingham)\
%D 1987
D MAG157 International Journal of General Systems\
%V 13\
%N 3\
%D 1987
------------------------------
Date: 9 Nov 87 16:57:20 GMT
From: honavar@speedy.wisc.edu (A Buggy AI Program)
Reply-to: honavar@speedy.wisc.edu (A Buggy AI Program)
Subject: Re: Success of AI
In article <4357@wisdom.BITNET> eitan%H@wiscvm.arpa (Eitan Shterenbaum) writes:
>
>As to the claim "the brain does it so why shouldn't the computer" -
>It seem to me that you forget that the brain is built slightly differently
>than a Von-Neuman machine ... It's a distributed enviorment lacking boolean
>algebra. I can hardly believe that even with all the partial solutions for
>all the complicated sets of NP problems that emulating a brain brings up, one
>might be able to present a working program. If you'd able to emulate mouse's
>brain you'd become a legend in your lifetime !
>Anyway, no one can emulate a system which has no specifications.
>if the neuro-biologists would present them then you'd have something to start
>with.
I use the term "computer" in a sense somewhat broader than a
Von-Neuman machine. We can, in principle, build machines that
incorporate distributed representations, processing and control.
It is not clear what you mean by a "distributed environment lacking
boolean algebra."
The use of fine-grained distributed representations naturally results
in behavior indicative of processes using fuzzy or probabilistic logic.
The goal is, not necessarily to emulate the brain in all its detail:
We can study birds to understand the principles of aerodynamics that
explain the phenomenon of flying and then go on to build an aeroplane
that is very different from a bird but still obeys the same laws of
physics. As for specifications, they can be provided in different
forms and at different levels of detail; Part of the exercise is
to discover such specifications - either by studying actual existing
systems or by analyzing the functions needed at an abstract level to
determine the basic building blocks and how they are to be put
together.
>
>And last - Computers aren't meta-capable machines they have constraints,
> not every problem has an answer and not every answermakes sense,
> NP problems are the best example.
>
Are you implying that humans are "meta-capable" - whatever that means?
VGH
------------------------------
Date: 10 Nov 1987 10:29-EST
From: Spencer.Star@B.GP.CS.CMU.EDU
Subject: Re: Guilding the Lemon
Something I was reading the other day may be of interest to those
involved in this discussion of doing a Ph.D. thesis that follows
closely someone else's work as opposed to striking off in some
completely new direction.
In Allen Newell's presidential address to AAAI in 1981, he comments on
the SIGART "Special Issue of Knowledge Representation" in which Ron
Brachman and Brian Smith present the answers to an elaborate
questionnaire sent to members of the AI community to find out their
views on knowledge representation.
"The main result was overwhelming diversity--a veritable jungle of
opinions. There is no consensus on any question of substance. ...
Many (but of course not all?) respondents themselves felt the same way.
As one said, 'Standard practice in the representation of knowledge is
the scandal of AI.'
"What is so overwhelming about the diversity is that it defies
characterization. ... There is no tidy space of underlying issues in
which respondents, hence the field, can be plotted to reveal a pattern
of concerns or issues. Not that Brachman and SMith could see. Not
that this reader could see."
By encouraging students to do their research on a subject by taking a
completely new approach, we are denying the value of previous work.
Certainly there is room for some Ph.D. students to take this path. But
a large part of what AI should be doing is building on the foundations
laid by the previous generations of researchers.
Spencer Star
------------------------------
Date: Mon, 9 Nov 87 15:11:19 PDT
From: ladkin@kestrel.ARPA (Peter Ladkin)
Subject: the wonder of words
gee, first ken laws says that maybe ai researchers don't need to think
too deeply, but maybe build whimsical experimental systems, and now
he's saying that automatic programming won't work because algorithms
are just too hard to design. i praise him for his consistency - one view
certainly follows from the other. i might use the old five-letter
expletive popularised by t.j. watson.
peter ladkin
ladkin@kestrel.arpa
------------------------------
Date: Tue, 10 Nov 87 11:25:52 MET
From: Laurent Siklossy <mcvax!cs.vu.nl!siklossy@uunet.UU.NET>
Subject: In Defense of FORTRAN
FORTRAN and other "standard" programming languages have
been used for years for advanced AI. One of the French AI
pioneers (if not THE pioneer, Ph.D. around 1961(?)),
Dr. Jacques Pitrat, has programmed for years in FORTRAN
with his own extensions. His programs included
discovering interesting logical theorems, learning in
the domain of games (chess), and many other areas.
Prof. Jean-Louis Lauriere wrote his Ph.D. thesis
(Universite de Paris VI, 1976; see his 100+ pages
article about that in the AI Journal, 1977 I think) in
PL/1. Lauriere's system was, in my opinion, the first
real (powerful) general problem solver, and remains a top
performing system in the field. (Lauriere may have been
pushed into using PL/1 by lack of other more appealing
choices, I cannot remember for sure.)
So it has been done, therefore you can do it too. I would
not recommend it, but that may be a matter of taste or
of limitations.
Laurent Siklossy
Free University, Amsterdam
siklossy@cs.vu.nl
---------------------------------------------------
Ken:
You are welcome to send above via the net if you find
it useful.
Cheers, LS
------------------------------
Date: Wed 11 Nov 87 21:42:50-PST
From: Laurence I. Press <LPRESS@venera.isi.edu>
Subject: FORTRANecdote
As a student assistant to Earl Hunt in the mid 1960s I wrote "concept
acquisition" programs in FORTRAN -- see the book Experiments in Induction,
Hunt, Marin and Stone, Academic Press, around 1965 if you don't believe it.
After that I wrote induction programs in JOVIAL too.
Larry
------------------------------
End of AIList Digest
********************
