Copy Link
Add to Bookmark
Report
AIList Digest Volume 1 Issue 004
AIList Digest Monday, 16 May 1983 Volume 1 : Issue 4
Today's Topics:
Research Posts in AI at Edinburgh University
AI at the AAAS
Expert System for IC Processiong
Re: Artificial languages
Loglan
Excerpt about AI from a NYTimes interview with Stanislaw Lem
----------------------------------------------------------------------
Date: Wednesday, 11-May-83 16:29:52-BST
From: DAVE FHL (on ERCC DEC-10) <bowen@edxa>
Reply-to: bowen%edxa%ucl-cs@isid
Subject: Research Posts in AI at Edinburgh University
--------
UNIVERSITY OF EDINBURGH
DEPARTMENT OF ARTIFICIAL INTELLIGENCE
2 Research Fellows
Applications are invited for these SERC-funded posts, tenable from
July 1 1983 or a mutually agreed date, to work on a project to
formulate techniques whereby an intelligent knowledge-based training
system can deduce what a user's aims are. Experience of UNIX and
programming is essential. Experience of PROLOG or LISP and some
knowledge of IKBS techniques would be an advantage. The posts are
tenable for three years, on the R1A scale (6375-11105 pounds).
Candidates should have a higher degree in a relevant discipline, such
as Mathematics, Computer Science or Experimental Psychology.
Applications, including a curriculum vitae and names of two referees,
should be sent to The Secretary's Office, University of Edinburgh, Old
College, South Bridge, Edinburgh EH8 9YL, Scotland, (from whom further
details can be obtained), by 31 May 1983.
------------------------------
Date: 13 May 83 10:53:04 EDT
From: DAVID.LEWIN <LEWIN@CMU-CS-C>
Subject: AI at the AAAS
The following session at the upcoming AAAS meeting should be of
interest to readers of AI-LIST.
ARTIFICIAL INTELLIGENCE: ITS SCIENCE AND APPLICATION
American Association for the Advancement of Science
Annual Meeting- Detroit, MI; Sunday, May 29, 1983
Arranged by: Daniel Berg, Provost--Carnegie-Mellon University
Raj Reddy, Director--Robotics Institute, CMU
"Robust Man-Machine Communication"
Jaime Carbonell, CMU
"Artificial Intelligence Applications in Electronic Manufacturing"
Samuel H. Fuller, Digital Equipment Corp. (Hudson, MA)
"Expert Systems in VLSI Design"
Mark Stefik, Xerox-PARC
"Science Needs in Artificial Intelligence"
Nils Nilsson, SRI International
"Medical Applicationns of Artificial Intelligence"
Jack D. Myers, Univ. of Pittsburgh
"The Application of Strategic Planning and Artificial Inntelligence to
the Management of the Urban Infrastructure"
Charles Steger, Virginia Polytechnic Inst. & State Univ.
------------------------------
Date: 14 May 1983 2154-PDT (Saturday)
From: ricks%UCBCAD@Berkeley
Subject: Expert System for IC Processiong
I'm about to start preliminary work on an expert system for integrated
circuit processing. At this time, its not clear whether it will deal
with diagnosing and correcting problems in a process line, or with
designing new process lines.
I would like to know if anybody has done any work in this area, and
what the readers of this list think about building an expert system
for this purpose.
I realize that this letter is somewhat vague, but I'm in the early
stages of this and I'd like to see what has been done and what options
I have.
Thanks,
Rick L Spickelmier
ricks@berkeley
University of California
Electronics Research Laboratory
Cory Hall
Berkeley, CA 94720
(415) 642-8186
------------------------------
Date: 11 May 1983 19:10 EDT
From: Stephen G. Rowley <SGR @ MIT-MC>
Subject: Artificial languages
Since people seem to be interested in artificial languages and the
Whorfian hypothesis, some words about Loglan might be interesting.
(If that's what started the discussion and I missed it, apologies to
all...)
Loglan is a language invented by J. Brown in the mid-50's to test the
Whorfian hypothesis with a radically different language. It's got a
simple grammar believed to be utterly unambiguous, a syntax based on
predicate calculus, and morpholgy that tells you what "part of speech"
(to stretch a term) a word is from its vowel-consonant pattern.
Of the 14 non-vacuous logical connectives, all are pronounceable in
one syllable. By comparison, English Dances about a LOT to say some
of them.
There are some books about it, and even a couple of regular journals.
Once upon a time, there was a Loglan mailing list here at MIT, but it
died of lack of interest.
-SGR
------------------------------
[Here is further info on Loglan culled from Human-Nets. -- KIL]
Date: 11 Dec 1981 2314-PST
From: JSP at WASHINGTON
Subject: Loglan as command language.
English is optimized to serve as a verbal means of communication
between intelligences. It would be highly surprising if it turned out
to be optimal for the much different task of communicating between an
intelligent (human) and a stupid (computer) via keyboard. In fact, it
would be surprising if English proved well suited to any sort of
precise description, given that various mathematical notations, Algol
and BNF, for example, all originated as attempts to escape the
ambiguity and opacity of English. (Correct me if I'm wrong, but I
seem to recall that Algol was originally a publication language for
human-human communication, programming applications coming later.)
Much the same may be said, with less force, for Loglan, which is
also targeted on human-human communication, albeit with a special
focus on simplicity and avoidance of syntactic ambiguity. (Other
Loglanists might not agree.)
For those interested, the Loglan Institute is alive and well, if
rather hard to find pending completion of a revised grammar and word
morphology. I'd be happy to correspond with anyone interested in the
language... and delighted to hear from any YACCaholic TL subscribers
interested in working on the grammar...
--Jeff Prothero
------------------------------
Date: 11 Dec 1981 06:46:30-PST
From: decvax!pur-ee!purdue!kad at Berkeley (Ken Dickey at Purdue CS)
Subject: Loglan
I have received several requests for more information on Loglan, a
language which may be ideal for man-computer communication. Here is a
brief description:
Synopsis: (from the book jacket of LOGLAN 1: A LOGICAL LANGUAGE, James
C. Brown, Third Edition)
Loglan is a language designed to test the Sapir-Whorf
hypothesis that the natural languages limit human thought. It does
this so by pushing those limits outward in predictable directions by:
*incorporating the notational elegance of symbolic logic (it is
TRANSFORMATIONALLY POWERFUL);
*forcing the fewest possible assumptions about "reality" on its
speakers (it is METAPHYSICALLY PARSIMONIOUS);
*removing all structural sources of ambiguity (in Loglon anything, no
matter how implausible, can be said clearly; for it is SYNTACTICALLY
UNAMBIGUOUS);
*generalizing all semantic operations (whatever can be done to any
Loglan word can be done to every Loglan word; for it is SEMANTICALLY
NON-RESTRICTIVE);
*deriving its basic word-stock from eight natural languages, including
three Oriental ones (it is therefore CULTURALLY NEUTRAL);
Notes:
Loglan has a small grammar (an order of magnitude smaller than
any "natural" grammar).
It is isomorphic (spelled phonetically-- all punctuation is
spoken).
There are a set of rules for word usage so that words are
uniquely resolvable (No "Marzee Dotes" problem).
The most frequently used grammatical operators are the
shortest words.
The word stock is derived from eight languages (Hindi,
Japanese, Mandarin Chinese, English, Spanish, Russian, French, and
German), weighted by usage for recognizability. I.e. using Loglan
rules to satisfy form, words are made up to be mnemonic to most of the
worlds speakers.
Loglan "predicates" are, in a sense, complete. For example
MATMA means X is the MOTHER of Y by father W. Joan matma == Joan is
the mother of .. by .. == Joan is a mother. Matma Paul == Paul's
mother, etc. These "slots" can change positions by means of
operators.
Modifiers precede modified words. Garfs school => a garfs
type of school (a school FOR garfs) as opposed to a school BELONGING
to garfs.
Language assumptions can be quite different. For example,
there are a number of words for "yes", meaning "yes, I will", "yes, I
agree", etc.
Although considered an experimental tool, there are people
that actually speak it. (It is a USEFUL tool).
Pointer: The Loglan Institute
2261 Soledad Rancho Road
San Diego, California 92109
As I am an armchair linguist, you should reference the above pointer
for more information.
-Ken
------------------------------
Date: 8 Apr 1982 01:32:44-PST
From: ihnss!houxi!u1100a!rick at Berkeley
Subject: Loglan
A while ago somebody (I believe it was in fa.human-nets during a
discussion of sexism in personal pronouns) asked the question "What
does Loglan do about gender?".
As usual with such questions the answer is not easy to describe in a
few words. But to simplify somewhat, Loglan has no concept of
grammatical gender at all. The language has a series of five words
that act (approximately) like third person pronouns, but there is no
notion of sex associated with them.
Loglan also does away with most of the usual grammatical categories,
such as "nouns", "adjectives" and "verbs". In their place it has a
single category called "predicate". Thus the loglan word "blanu" can
be variously translated as "blue" (an adjective), "is a blue thing" (a
verb-like usage), and "blue thing" (a noun-like usage).
Loglan is uninflected. It has no declensions or conjugations. But it
does have a flock of "little words" that serve various grammatical and
punctuational purposes. They also take the place of such affixes as
"-ness" (as in "blueness") in English.
More information about Loglan can be gotten by writing to:
The Loglan Institute, Inc.
2261 Soledad Rancho Road
San Diego, CA 92109
------------------------------
Date: Sun 15 May 83 12:17:41-PDT
From: Robert Amsler <AMSLER@SRI-AI.ARPA>
Subject: Excerpt about AI from a NYTimes interview with Stanislaw Lem
Sunday, March 20th, NYTimes Book Review Interview with Stanislaw Lem
by Peter Engel
Interviewer: "You mentioned robots, and certainly one of the most
important themes in your writing is the equality of men and robots as
thinking, sentient beings. Do you feel that artificial intelligence
at this level will be achieved within the forseeable future?"
Lem: "My opinion is that in roughly 100 years we will arrive at an
artificial intelligence that is more intelligent and reasonable than
human intelligence, but it will be completely different. There are no
signs indicating that computers will in certain fields become equal to
men. You should not be misled by the fact that you can play chess with
a computer. If you want to accomplish certain individual tasks,
computers are fine. But when you are talking about psychological
matters, every one of us carries in his head the heritage of the
armored fish, the dinosaurs, and other mammals. These limitations do
not exist outside the domain of biological evolution. And there's no
reason why we should imitate them -- the very idea is silly. In the
field of mechanics it would be the same as if the Arabs were to say
they didn't want airplanes and automobiles, only improved camels. Or
that you shouldn't supply automobiles with wheels, that you must
invent mechanical legs.
I'm going to show you a book. 'Golem XIV' is going to be published
next year in America. It's a story about the construction of a
supercomputer and how it didn't want to solve the military task it was
given, the purpose it had been constructed for in the first place. So
it started to devote itself to higher philosophical problems. There
are two stories in 'Golem XIV,' two lectures for scientists. In the
first Golem talks about humans and the way it sees them, in the second
about itself. It tries to explain that it's already arrived at a level
of biological evolution will never reach on it own (sic). It's on the
lowest rung of a ladder, and above it there might exist now or in the
future more potent intelligences. Golem does not know whether there
are any bounds in its progress to the upper sphere. And when it, in a
manner of speaking, takes leave of man, it is primarily for the
purpose of advancing further up this ladder.
In my own view, man will probably never be able to understand and
recognize everything directly, but in an indirect manner he will be
able to achieve command of everything if he constructs intelligence
amplifiers to fulfill his wishes. Like a small child, he will then be
receiving gifts. But he will not be able to perceive the world
directly, like a small child who is given an electric railway. The
child can play with it, he can even dismantle it, but he will not
understand Maxwell's theory of electricity. The main difference is
that the child will one day become an adult, and then if he wants he
will eventually study and understand Maxwell's theory. But we will
never grow up any further. We will only be able to receive gifts from
the giants of intelligence that we'll be able to build. There is a
limit to human perception, and beyond this horizon the fruit of
observation will be gleaned from other beings, research machines or
whatever. Progress may continue, but we will somehow be staying
behind."
------------------------------
End of AIList Digest
********************
