Copy Link
Add to Bookmark
Report
AIList Digest Volume 4 Issue 115
AIList Digest Wednesday, 7 May 1986 Volume 4 : Issue 115
Today's Topics:
Query - Computer Resources Management,
Literature - Connection Machine Articles & U. Kansas Tech Reports,
Philosophy - Consciousness,
AI Approaches - Learning,
Linguistics - Trademarks
----------------------------------------------------------------------
Date: Mon 5 May 86 09:48:03-PDT
From: Jean-Pierre Dumas <DUMAS@SUMEX-AIM.ARPA>
Subject: Computer resources management
I post this for a friend.
I will forward any mail.
dumas@sumex
I am interested in computer system performance analysis modelling.
I am indeed developing a system, based on AI techniques, to deal with
tuning and performance planning of computer system management.
I would be delighted to be in touch with people concerned with this question.
Address :
Dr. Saddek BELAID
CISI-TELEMATIQUE
CEN-SACLAY BP 24
91190 GIF/YVETTE
FRANCE
Phone : (+33) 1 69 08 20 12
------------------------------
Date: 1 May 86 18:46:00 GMT
From: hplabs!hpfcdc!hpfcla!hpcnoe!jd@ucbvax.berkeley.edu
Subject: Re: connection machine articles
Daniel Hillis has written a book entitled
"The Connection Machine". It is available through the Library of Computer
and Informaiton Sciences (book club). I am sure that It can be found
elsewhere.I just recieved the book and It seems very readable if not
intriguing.
Hope I have Helped,
John Dye Hewlett Packard
{inhp4|hplabs}!hpfcla!hpcnoa!jd Colorado Networks Division
------------------------------
Date: 4 May 86 05:51:56 GMT
From: nike!topaz!harvard!think!bruce@ucbvax.berkeley.edu (Bruce J. Nemnich)
Subject: Re: connection machine articles
Also, if all you have read is the old AI memo, you should definitely
read the book, |The Connection Machine|, by Danny Hillis, published
last year by the MIT Press.
--
--Bruce Nemnich, Thinking Machines Corporation, Cambridge, MA
--bruce@think.com, ihnp4!think!bruce; +1 617 876 1111
------------------------------
Date: Fri, 2 May 86 10:16:26 CDT
From: Glenn Veach <veach%ukans.csnet@CSNET-RELAY.ARPA>
Subject: Tech reports - University of Kansas
Following is a list of recent technical reports which have
been issued by the department of Computer Science of the
University of Kansas in conjunction with research done in
the department's Artificial Intelligence Laboratory.
%A Frank M. Brown
%T Reasoning in a Hierarchy of Deontic Defaults
%I Department of Computer Science, University of Kansas
%R TR-86-2
%X A commonsense theory of reasoning is presented which models
our intuitive ability of reason about defaults involving both
deontic and doxastic logic. The concepts of this theory do not
involve fixed points or Kripke semantics but instead are explicitly
defined in a modal quantificational logic which captures the modal
notion of logical truth. An example involving derivations of
obligations from both a robot's beliefs and a hierarchy of deontic
defaults is given. To be published in the proceedings of the
1986 Canadian Artificial Intelligence Conference. 11 pp.
%A Frank M. Brown
%T Toward a Commonsense Theory of Nonmonotonicity
%I Department of Computer Science, University of Kansas
%R TR-86-3
%X A logical theory of nonmonotonic reasoning is presented which
permits a commonsense approach to defaults. The axioms and inference
rules for a modal logic based on the concept of logical truth are
described herein along with basic theorems about nonmonotonic
reasoning. An application to the frame problem in robot plan
formation is presented. To be published in the proceedings of the
Eight International Conference on Automated Deduction. 12 pp.
%A Frank M. Brown
%T A Comparison of the Commonsense and Fixed Point Theories
of Nonmonotonicity
%I Department of Computer Science, University of Kansas
%R TR-86-4
%X The mathematical fixed point theories of nonmonotonic reasoning
are examined and compared to a commonsense theory of nonmonotonic
reasoning which models our intuitive ability to reason about defaults.
It is shown that all of the known problems of the fixed point theories
are solved by the commonsense theory. The concepts of this commonsense
theory do not involve mathematical fixed points, but instead are
explicitly defined in a monotonic modal quantificational logic which
captures the modal notion of logical truth. 12 pp.
%A Frank M. Brown
%T An Experimental Logic Based on the Fundamental Deduction Principle
%I Department of Computer Science, University of Kansas
%R TR-86-5
%X Experimental logic can be viewed as a branch of logic dealing with
the actual construction of useful deductive systems and their application
to various scientific disciplines. In this paper we describe an
experimental deductive system called the SYMbolic EVALuator (i.e. SYMEVAL)
which is based on a rather simple, yet startling principle about deduction,
namely that deduction is fundamentally a process of replacing expressions
by logically equivalent expressions. This principle applies both to
logical and domain dependent axioms and rules. Unlike more well known
logical inference systems which do not satisfy this principle, herein is
described a system of logical axioms and rules called the SYMMETRIC LOGIC
which is based on this principle. Evidence for this principle is given
by proving theorems and performing deduction in the areas of set theory,
logic programming, natural language analysis, program verification,
automatic complexity analysis, and inductive reasoning. To be published
in the international journal Artificial Intelligence. 120 pp.
%A Frank M. Brown
%T Automatic Deduction in Set Theory
%I Department of Computer Science, University of Kansas
%R TR-86-6
%X A proof of the definability of ordered pairs in set theory is described
and discussed. This proof was obtained in an entirely automatic way using
the SYMEVAL deduction system and the SYMMETRIC LOGIC axioms. The analogous
point in this proof where other theorem proving methods and systems have
failed to prove this theorem are described. The ability of this system to
automatically derive one half of this theorem from the other half is also
discussed, thus showing that this kind of deduction system can be used to
produce answers other than just yes/no answers to mathematical questions.
24 pp.
%A Frank M. Brown
%T An Experimental Logic
%I Department of Computer Science, University of Kansas
%R TR-86-7
%X The fundamental deduction principle, SYMEVAL deductive system, and
SYMMETRIC LOGIC are introduced. Theorems are proved in the area of
set theory, complexity analysis and program verification. 17 pp.
%A Frank M. Brown
%T Logic Programming with an Experimental Logic
%I Department of Computer Science, University of Kansas
%R TR-86-8
%X In this paper we describe the experimental programming logic which
uses the deductive system SYMEVAL which is based on the fundamental
deduction principle. Theorems and deductions are performed in the area
of logic programming and then discussed as they relate to the above
principle. 18 pp.
------------------------------
Date: 29 Apr 86 16:19:45 GMT
From: tektronix!uw-beaver!fluke!ssc-vax!bcsaic!michaelm@ucbvax.berkeley
.edu (michael maxwell)
Subject: Consciousness
In light of the recent flurry of articles on consciousness (of computers,
toasters, Endomoeba histolitica...), some readers may be interested in a
recent book: "Animal Thinking", by Donald Griffin. I've just started reading
it, so I can't say much; but the author is an ethologist (=student of animal
behavior), and his contention is that many animals *are* concious.
--
Mike Maxwell
Boeing Artificial Intelligence Center
...uw-beaver!uw-june!bcsaic!michaelm
------------------------------
Date: 29 Apr 86 13:26:18 GMT
From: ucdavis!lll-lcc!lll-crg!seismo!mcvax!ukc!warwick!gordon@ucbvax.
berkeley.edu (Gordon Joly)
Subject: Plan 5 for Inner Space - A Sense of Mind.
The following project has been proposed. Design and implement an
Intelligent System with the following characteristics. The system
is to run in real time and be portable. Memory and processing units
must be in the same enclosure, run on low power and not overheat.
(a) Full colour binocular vision, with motion perception.
(b) Speech processing and speech synthesis.
(c) Natural language ability :-
(1) Semantic ability
(2) Translation.
(3) Ability to summarise.
(4) Humour (optional).
(d) Learning ability.
(e) Ability to control a large number of servo mechanisms, with strength
and sensitivity.
(f) Other tasks, as yet unspecified, but the system must be able
to cope with extra requirements, as and when the need arises,
using characteristic (d).
Queries: Time to completion? Cost?
------------------------------
Date: 5 May 86 11:33:00 MST
From: fritts@afotec
Reply-to: <fritts@afotec>
Subject: Re: Science 86 Article on Neural Nets
I N T E R O F F I C E M E M O R A N D U M
Date: 5-May-1986 10:18 MDT
From: STEVE E. FRITTS
FRITTS
Dept: 1859ISSS/SIP
Tel: 505-846-2595
TO: Remote Addressee ( _MAILER!AILIST@SRI-AI )
Subject: Re: Science 86 Article on Neural Nets
Most of what I've read on this list appears to place AI closer to the
"Frankenstein" theory of assembling intelligence, fully formed and
functioning, like any other computer program; just push the button and
watch it go.
Neural networks appear to be more natural in their approach. Terry
Sejnowski's NETalk was equipped more with rules on how to learn to
perform a task than rules on how to perform a task. I think that this
is a crucial difference. If a computer is programmed only to perform
a task, then the programmer must design for every possible eventuality
which may affect the performance and for every possible consequence or
outcome. The problem is that such a program, no matter how
comprehensive, makes assumptions. These assumptions are fatal for
intelligence. They doom the program as surely as evolution dooms some
species of life on this planet, and for much the same reason. Human
intelligence may have developed as the ultimate weapon against
changing environments; better than adaptation because it allows for
greater variety of response.
So, if "intelligence" is developed in a computer through learning
mechanisms rather than assembled by means of cunning rules and
algorithms, perhaps it stands a better chance of achieving sufficient
universality that it may compete with the human mind. Odd that we
would dream of building our own competition.
I vaguely recall that a long time ago there were machines called analog
computers which worked on a principle of varying voltages and
resistances rather than the digital machine's method of detecting the
polarity (the "on" or "off" state) of a particular circuit junction.
Hopfield and Tank's neural net appears to perform in some ways similar
to an analog computer. The article is too general on the technicals
of a "neural net" machine and I add my request to others on this list
for a little better technical description. Also, perhaps someone will
enlighten me about the possible relevance or irrelevance of analog
computers to neural nets.
DISCLAIMER: My opinions are my own alone and do not represent any
official position by my employer.
Steve Fritts
FRITTS@AFOTEC.ARPA
------------------------------
Date: Fri 2 May 86 11:33:51-CDT
From: Gordon Novak Jr. <AI.NOVAK@R20.UTEXAS.EDU>
Subject: Xerox as a verb
I think it was William Safire who stated the metarule:
You can verb anything.
At least in Washington.
------------------------------
Date: 2 May 86 13:30 PDT
From: Stern.pasa@Xerox.COM
Subject: Trademarks
There's language and language. Yes, there is a lexical distinction
between xerox and Xerox, but does the law recognize lexical vs other
discriminations between usages? Yes, sort of, maybe.
One might use syntactic or semantic analysis of usages to determine
whether it is acceptable to use a trademarked word, but NL workers know
how difficult it is to produce an absolute semantic analysis of
unrestricted language.
So instead, companies protecting trademarks work at a higher level of
abstraction, the real-world script. They know that any unrestricted use
of a word leads to bad consequences, so their heuristic for when to
complain is based on factors exogenous to the apparent linguistic
content in which the reference appears.
In conclusion it is fascinating that the legal protection of trademarks
should involve such a wide variety of linguistic considerations. There
are no easy answers to the theoretical questions involved, but as KIL
points out, there are laws.
P.S. None of the above is to be taken as the position of my employer.
Josh
------------------------------
Date: Fri 2 May 86 17:24:26-PDT
From: Pat Hayes <PHayes@SRI-KL>
Subject: Re: Names and Trademarks
Of course, I understand that the company is legally obliged to correct
uses of its name in order to retain trademark status. My point was that
as a matter of fact, like it or not, to xerox is now a verb in common
usage, and is going to stay that way. I have a dictionary ( Random
House ) in which Xerox(TM) is both a noun and a verb, and xerography (
no capitalisation or trademark ) is another noun. Clearly, the company
insisted that the publishers gave trademark acknowledgement: equally
clearly, both the publishers and the company acknowledge that the word is
part of the language.
There are important differences between xerox and asprin on one hand, and
exxon and frigidaire and IBM on the other. In these latter cases, identification
was due to the company having a dominating position in the market, and
nothing else. In the former cases, it was the only owner and supplier of
a vital piece of technology, during the period in which it became a
commonplace of everyday life, and indeed transformed everyday life. And
finally, a last difference is that in both the former cases there was and
is no alternative way of referring to the things available. What is a Canon
xerox machine if it isnt that? It has to be some awkward neologism like
a dry copier, or a copier using the xerographic process. What would we call
an asprin if we couldnt call it an asprin? Like Bayer, you guys have been
too successful: not only did you invent the process, you also invented the
way of talking about it.
Patrick J. Hayes (TM)
PS> I am told that in the oil business, to schlumberger is a common verb.
------------------------------
Date: Sat 3 May 86 02:20:42-PDT
From: Lee Altenberg <ALTENBERG@SUMEX-AIM.ARPA>
Subject: "Exxon" vs. "exon" .
I remember reading in Time Magazine around 1970 when ENCO/ESSO Oil Corp.
unveiled its new name, "Exxon". They had looked through all the major
languages of the world to find one word that didn't mean anything in any of
them. "Exxon" was what they found.
Alas. In 1978, Walter Gilbert of Harvard came up with new words
for the newly discovered pieces that eukaryotic genes are made of: "intron"
for the spliced out sequences, and "exon" for the expressed sequences.
I wonder what went through the minds of the CEOs of Exxon when they caught
wind of this?
I should mention a sign I saw in the Bioengineering Dept. at UC
Berkeley:
--AXXON--
MOTOR NEURON SERVICE
-Lee Altenberg
------------------------------
End of AIList Digest
********************
