Copy Link
Add to Bookmark
Report
AIList Digest Volume 2 Issue 024
AIList Digest Tuesday, 6 Mar 1984 Volume 2 : Issue 24
Today's Topics:
Conferences - AAAI-84 Paper Submission Deadline,
AI Tools - LISP for IBM PC & UNIX VAX Tools,
Manual Generators - Replys,
Parser Generator - Request,
Mathematics - Fermat's Last Theorem & Map Coloring,
Personal Robotics - Reply,
Waveform Analysis - ECG Systems & Validation,
Review - U.S. Response to Japan's AI efforts
----------------------------------------------------------------------
Date: Wed 29 Feb 84 15:44:06-PST
From: Ron Brachman <Brachman at SRI-KL>
Subject: AAAI-84 Paper Submission Deadline
******* AAAI-84 PAPER SUBMISSION DEADLINE IS APRIL 2, 1984 *******
The SIGART Newsletter (No. 87, January 1984) has mistakenly published
two conflicting dates for submission of papers to AAAI-84. Please note
that papers must be received in the AAAI Office in Menlo Park, CA, on or
before April 2, 1984. This is the date that appears in the AAAI-84 Call
for Papers (printed on page 17 of the above-mentioned Newsletter). The
date printed in the "Calendar" section on page 1 of the Newsletter is
incorrect.
Thank you,
Ron Brachman, Program Chair
Claudia Mazzetti, AAAI Executive Director
------------------------------
Date: Sun 4 Mar 84 13:33:49-PST
From: Ted Markowitz <G.TJM@SU-SCORE.ARPA>
Subject: LISP for IBM PC
I asked the list a while back about implementations of LISPs for
IBM PC's. I got a pointer for IQLISP, but seem to have misplaced
the pertinent info on how to order it. Can anyone supply this?
If you have any other implementations, I'll be glad to pass any
reviews back to the list.
--ted
[The original message must have been prior to issue 53, and I
don't have it online. Does some have the address handy? -- KIL]
------------------------------
Date: 27 Feb 84 16:26:42-PST (Mon)
From: ihnp4!houxm!hou2a!zev @ Ucb-Vax
Subject: AI (LISP,PROLOG,ETC.) for UNIX VAX
Article-I.D.: hou2a.269
A friend of mine is looking for a LISP, PROLOG, and/or
any other decent Artificial Intelligence system that
will run on a VAX under UNIX.
Please send replies directly to Mr. Leonard Brandwein at
aecom!brandw
He asked me to post this as a favor, since he does not
have direct access to the net.
In the likely case that you don't have a direct path
to aecom, here is one that will get you there from
any machine that can reach houxm:
houxm!hou2a!allegra!philabs!aecom!brandw
Of course, you can shorten the path if you can reach
any of the intermediate machines directly.
Thank you very much.
Zev Farkas hou2a!zev 201 949 3821
[When sending to Usenet from the Arpanet, be sure to put double quotes
around all of the address prior to the @-sign. Readers who want help
getting messages through the gateways should contact AIList-Request@SRI-AI.
Useful summaries or interesting replys may be published directly in
AIList, of course. I will pass along some information about CProlog
in the next issue. -- KIL]
------------------------------
Date: Thu, 1 Mar 84 5:12:55 EST
From: Stephen Wolff <steve@brl-bmd>
Subject: Re: AI (LISP,PROLOG,ETC.) for UNIX VAX
[Forwarded from the Info-Unix distribution by Laws@SRI-AI.]
Franz Lisp comes with Berkeley UNIX. Interlisp is available. Also T.
CProlog is available from Edinburgh. You can get Rosie from RAND.
And these are just basics. There's LOTS! There are many schools out there
who are (possibly newly) in the AI business who couldn't afford DEC-20's
(obviously not SRI, UTexas, CMU, etc.), but who DID buy VAXen back when they
were good value for money. And they're mostly running BSD, and they're
busily developing all the tools and software that AI folk do. Is there any
PARTICULAR branch of AI you're interested in? [...]
------------------------------
Date: Thu, 1 Mar 84 4:23:11 EST
From: Stephen Wolff <steve@brl-bmd>
Subject: Documentation tools
Artificially intelligent it's not, and not even fancy; but there are
folks hereabouts that use the UNIX tools SCCS (or RCS) to do documentation
of various sorts. Although intended for managing the writing, evolving and
maintaining of large software packages, they can't tell C from Fortran from
straight English text and they will quite cheerfully maintain for you the
update/revision tree in any case.
I should imagine with a bit if thought you could link your code AND
documentation modules and manage 'em both simultaneously and equitably.
------------------------------
Date: Sat, 3 Mar 84 18:43:59 PST
From: Charlie Crummer <crummer@AEROSPACE>
Subject: Manual generators
The SCRIBE system (Brian K. Reid of CMU and Janet H. Walker of BBN)
may be close to what you are looking for. It has automatic paragraph
numbering, automatic table-of-contents generation, automatic indexing,
and automatic bibliography. (I use the word "automatic" somewhat
loosely. The user has to be involved.) A more sophisticated system,
I believe, is in use at the University of Michigan's Information
Systems Design and Optimization System (ISDOS) project. The contact
is Prof. Dan Teichroew in the Industrial and Operations Engineering
department at Ann Arbor. It may be avaliable to ISDOS sponsors.
--Charlie
------------------------------
Date: Thu 1 Mar 84 20:24:33-EST
From: Howard Reubenstein <HBR@MIT-XX.ARPA>
Subject: Looking for a Parser Generator
[Forwarded from the MIT-MC bboard by SASW@MIT-MC.]
A friend of mine needs a parser generator which produces
output in either FORTRAN or LISP. Does anyone know where he can
get access to one?
------------------------------
Date: Thu, 1 Mar 84 08:34 EST
From: MJackson.Wbst@PARC-MAXC.ARPA
Subject: Re: Fermat's Last Theorem & Undecidable Propositions
Fermat's Last Theorem:
is the assertion that
A^N + B^N = C^N
has no solution in integers for N > 2. (For N = 2, of course, all the
well-known right triangles like [3,4,5] are solutions.)
The Four-Color Theorem:
states that any planar map can be colored so that no two adjacent
regions are the same color using no more than four different colors.
(Regions must be connected; "adjacent" means having a common boundary of
finite length, i.e. not just touching at a point.
The latter was shown to be true by two mathematicians at the University
of Illinois, using a combination of traditional mathematical reasoning
and computer-assisted analysis of a large set of graphs. An article
describing the proof can be found in a back issue of /Scientific
American/.
The former appears in a manuscript by Fermat, with a marginal notation
to the effect that he had found a slick proof, but didn't have enough
space to write it down. This was discovered after his death, of course.
Most mathematicians believe the theorem to be true, and most do not
think Fermat is likely to have found a valid proof, but neither
proposition has been proved beyond question.
Mark
------------------------------
Date: 28 Feb 84 20:42:40-PST (Tue)
From: decvax!genrad!wjh12!n44a!ima!inmet!andrew @ Ucb-Vax
Subject: Re: Re: Fermat's Last Theorem & Undecida - (nf)
Article-I.D.: inmet.945
Fermat's Last Theorem states that the equation
n n n
A + B = C
has solutions in positive integers a, b, c, n only when n = 2.
The "four-color map problem" states that any map (think of, say, a map of the
US) requires at most four colors to color all regions without using the same
color for any two adjacent ones. (This is for 2-dimensional maps. Maps
on a sphere or torus require more - 5 and 7, I think.)
The former has neither been proven nor disproven. The latter was "proven"
with the aid of a computer program; many feel that this does not constitute
a true proof (see all the flames elsewhere in this group). Incidentally,
the school where it was "proven" changed their postage meters to print
"FOUR COLORS SUFFICE" on outgoing mail.
------------------------------
Date: Thu 1 Mar 84 13:53:05-PST
From: Sam Hahn <SHahn@SUMEX-AIM.ARPA>
Subject: Domestic Robotics
I find that Robotics Age (the journal of intelligent machines), published by
Robotics Age, Inc, located at:
Strand Building
174 Concord Street
Peterborough, NH 03458 (603) 924-7136
is a good source of information on low-end, more personal, and thus more
"domestic"ly oriented robotics. For example, the advertisers include
Micromation: voice command system for Hero-1
Iowa Precision Robotics:
68000-controlled educ/pers'l robot
Micron Techn.: computer vision for your PC
S.M. Robotics: PR kit for $59.95
just to name a few from the February 1984 issue.
Their articles are also more PR-oriented, and often include some level of
design info.
I'm new to the publication myself (about 1/2 year), but find it a source of
information not elsewhere available.
-- sam hahn
------------------------------
Date: 27 Feb 84 19:25:34-PST (Mon)
From: pur-ee!uiucdcs!marcel @ Ucb-Vax
Subject: Re: computer ECG - (nf)
Article-I.D.: uiucdcs.5890
Ivan Bratko, of the Josef Stefan Institute in Ljubljana, Yugoslavia, has
recently achieved some remarkable results. With the aid of computer
simulation he has built an expert system capable of diagnosing multiple
simultaneous heart malfunction causes from ECG outputs. This was a
significant contribution to medical science, since for the class of failures
he treated, there was no known method of diagnosing anything more complicated
than a single cause.
His work will be printed as a monograph from the newly-formed "International
School for the Synthesis of Expert Knowledge" (ISSEK), which will have its
first meeting this summer. ISSEK is an affiliation of computer science labs
dedicated to the automatic generation of new knowledge of super-human quality.
(Membership of ISSEK is by invitation only).
Marcel Schoppers
U of Illinois @ Urbana-Champaign
{ pur-ee | ihnp4 } ! uiucdcs ! marcel
------------------------------
Date: 2 Mar 84 20:54:42 EST
From: Ron <FISCHER@RUTGERS.ARPA>
Subject: Re: computer ECG, FDA testing of AI programs
Apparently because of fierce competition, much current information,
particularly with regard to algorithms, is proprietary. Worst in this
regard (a purely personal opinion) is HP who seems to think nobody but
HP needs to know how they do things and physicians are too dumb to
understand anyway.
...
They offer an advantage to small hospitals by offering verification of
the analysis by a Cardiologist (for an extra fee).
What the latter seems to say is that the responsibility for accepting
the diagnosis is that of the local cardiologist. I cannot see a
responsable doctor examining a few runs of a program's output and
proclaiming it "correct."
A hedge against complaints of computers taking over decision making
processes from human has been that we can look at the algorithms
ourselves or examine the reasons that a system concluded something.
If this information becomes proprietary the government will probably
license software for medical purposes the way the FDA does for new
drugs.
Imagine a testing procedure for medical diagnostic AI programs that is
as expensive and complicated as that for testing new drugs.
(ron)
[Ron makes a good point. As a side issue, though, I would like
to mention that H-P has not been entirely secretive about its
techniques. On March 8, Jim Lindauer of H-P will present a seminar
at Stanford (MJH 352, 2:45PM) on "Uses of Decision Trees in ECG
Analysis". -- KIL]
------------------------------
Date: 29 Feb 84 15:36:33 PST (Wednesday)
From: Hoffman.es@PARC-MAXC.ARPA
Subject: U.S. Response to Japan's AI efforts
In the new "soft" computer journal from Springer-Verlag, 'Abacus', Vol.
1, #2, Winter 1984, is an essay by Eric A. Weiss reviewing Feigenbaum
and McCorduck's 'Fifth Generation' book and general AI books. The
general A.I. review is worth reading. The whole piece is lengthy, but I
quote only from the final section.
--Rodney Hoffman
[This is a rather remarkable book review. In addition to discussing the
"The Fifth Generation" and several AI reference works and textbooks,
Eric Weiss describes the history and current partitioning of AI, the
disputes and alignments of the major AI centers, and the solution to
our technological race with foreign powers. It's well worth reading.
This second issue of Abacus also has interesting articles on Ada,
the language and the countess, tomographic and NMR imaging (with
equations!), and the U.S. vs. IBM antitrust suit, as well as columns
on computers and the laws and other topics. The magazine resembles
a Scientific American for the computer-oriented, and the NMR article
is of quality comparable to IEEE Computer. -- KIL]
------------------------------------------------------
U.S. Response
On the basis of all this perspective, let me return to the Fifth
Generation Project itself and suggest that the U.S. response should be
thoughtful, considered, not guided by panic or fear, but based on
principles this nation has found fruitful:
build on experience
do what you do best
encourage enthusiasm
What has been our experience with foreign science and technology? We
know that new scientific knowledge gives the greatest benefit to those
nations which are most ready to exploit and use it, and this ready group
may not include the originating nation.... [discussion of rocketry,
automobiles, shipbuilding, steel, consumer electronics]
From this experience, the U.S. should look forward to reaping the
benefits from whatever the Japanese Fifth Generation Project develops,
and, just because we are bigger, richer, and stronger, benefiting more
from these improvements than the originating nation....
... "Do what you do best." We do not compete with the Japanese very
well, but we do best in helping them.... [The U.S.] is best at helping
others, especially Japan, and at giving money away.... Thus, the
indicated course for the U.S. ... is to help the Japanese Fifth
Generation Project in every way we can: by supplying grants of money; by
loaning college professors; by buying and copying its product,
exploiting its scientific and technological developments and
breakthroughs as fast as they appear; and by ignoring or clucking
sympathetically over any failures or missed schedules. Finally,...
encourage enthusiasm.
Young military people may murmur against this stance on the grounds that
military developments must be home-grown and that the development of
technology which might be used in weapons should be guided by the
military. This assertion is borne out neither by history nor by the
present public attitude of the DoD.... [discussion of WWII anti-aircraft
guns, mines, torpedoes, and many other such]
... The advantages of letting another nation develop your military
hardware are frequently and forcefully explained to other countries by
the DoD and its industrial toadies, but these logical arguments... are
never put in their equally logical vice-versa form....
The danger is not that the Japanese will succeed -- for their successes
will result in U.S. benefits -- but that somehow we will not make prompt
use of whatever they accomplish. We might manage this neglect if we
overdo our national inclination to fight them and compete with them....
A related but more serious danger lies in the possibility that our
military people will get their thumbs into the American AI efforts and
make secret whatever they don't gum up.... Even the best ideas can be
killed, hurt, or at least delayed if hedged around with bureaucrats and
secrecy limitations.
... We should press vigorously forward on all fronts in the unplanned
and uncoordinated fashion that we all understand. We should let a
thousand flowers bloom. We should encourage everyone.... We should hand
out money. We should transport experts. We should jump up and down.
We should be ready to grab anybody's invention , even our own, and use
it. We should be ready to seize winners and dump losers, even our own.
We should look big, fearless, happy, and greedy, and not tiny,
frightened, worried, and dumb.
... The conclusion is: don't bet on the Japanese, don't bet against
them, don't fear them. Push forward with confidence that the U.S. will
muddle through -- if it can keep its government from making magnificent
plans for everyone.
------------------------------
End of AIList Digest
********************
