Copy Link
Add to Bookmark
Report
AIList Digest Volume 4 Issue 158
AIList Digest Thursday, 26 Jun 1986 Volume 4 : Issue 158
Today's Topics:
Literature - AIList in Technology Review & AI Expert,
AI Tools - Turbo Prolog & Language Paradigms,
Psychology - Memory in Bees & Creativity & Forward Following,
Policy - Covert Ads
----------------------------------------------------------------------
Date: Wed 25 Jun 86 16:43:41-PDT
From: Ken Laws <Laws@SRI-AI.ARPA>
Subject: Technology Review
AIList is the subject of the First Line column (by John Mattill) in
the May/June 1986 issue of MIT's Technology Review, p. 2. John is
commenting on our discussion of the Dreyfus' article in their January
issue, and finds our "electronic gossip" an intriguing publication
channel. He quotes Peter Ladkin and me, and also Brad Miller's
"In 3,000 years philosophy has still not lived up to its promises,
and there is no reason to think it ever will." (He unfortunately
lists it as anonymous.) The editorial is followed by a letters column,
from which I particularly liked A. DeLuca's comment: "Granted, the mind
is not like a computer. But an airplane is not like a bird, either."
-- Ken Laws
------------------------------
Date: Mon, 23 Jun 86 14:10:48 CDT
From: Glenn Veach <veach%ukans.csnet@CSNET-RELAY.ARPA>
Subject: Summarizing "AI Expert"
The following is a summary/personal comment, of the "Premier Issue" of
"AI Expert" presented as "The Magazine for the Artificial Intelligence
Community."
As the editor, Craig LaGrow, notes it is "rare to find such a strong
response from advertisers" and "quality submissions" of articles in
the first run of a new magazine (particularly in the computer field
with our overwhelming number of upstart publications).
The advisory board of the magazine is quite impressive with the likes of
John Seely Brown of Xerox PARC, Carl Hewitt of MIT, Earl Sacerdoti of
Teknowledge, Donald Waterman of Rand, and Terry Winograd of Stanford,
to name a few. With the commitment of such names one should expect to
to see a quality publication.
The feature articles as well as the "regular columns" are well written
and contributed by knowledgeable authors. A feature for which the publisher
should be commended, is the inclusion in many of the articles of actual code
which demonstrates the technique or program which the author is presenting.
By the way, the code is reportedly available for downloading from four
different sources, for those who wish to try it out.
Following is an annotated list of articles and columns:
"Brain Waves" a column by Larry Geisel/CEO Carnegie Group Inc.
in this article he writes on "The AI Explosion: A Response to National
Priorities"
"AI Insider" a column capsulizing industry and academic developments by
Susan J. Shepherd a consultant with Academy for Educational Development
"Expert's Toolbox" a column written by Jonathan Amsterdam a grad student
at MIT writes an article on "Augmented Transition Networks for Natural
Language Parsing" which includes code for an ATN compiler and a sample
ATN grammar.
"AI Apprentice" a column by Bill and Bev Thompson free-lance writers and
consultants write on "PROLOG From the Bottom Up" which introduces PROLOG
and the basic logical concepts, includes some basic coded procedures.
"Control Over Inexact Reasoning" a feature article by Koenraad Lecot a grad
student at UCLA and D. Stott Parker a prof. at UCLA.
"Concurrency in Intelligent Systems" a feature by Carl Hewitt of MIT.
"Rule-Based Programming in OPS83" a feature by Dan Neiman with ITT Advanced
Technology Center and John Martin of Philips Laboratories. Includes code for
a short program.
"Multitasking for Common LISP" by Andrew Bernat of the University of Texas
at El Paso. Includes code for the concurrent processing modules.
"In Practice" a column by Henry Eric Firdman a consultant uses this column
to look at the use of AI in real-world business applications. This issue's
article -- "Components of AI Systems".
"Software Review" a column by Darryl Rubin of Microsoft. Here we get a look
at "Turbo PROLOG: A PROLOG Compiler for the PC Programmer."
"Book Store" a column by Lance B. Eliot, director of UCLA's Expert Systems
Laboratory gives short blips on four "classics".
"AI Expert" will be published monthly beginning in October. The above is
a sample of what they have to offer at this time. If they continue to
produce similar articles it should be of interest to most in the AI community
and especially those in industry seeking to apply AI to their needs, as well
as to those just starting to "get into" the field.
Subscription info: AI Expert PO Box 10952 Palo Alto CA 94303-0968 charter
first year at $27.00 for the 12 issues.
------------------------------
Date: Sun, 22 Jun 86 12:54:16 mdt
From: ted%nmsu.csnet@CSNET-RELAY.ARPA
Subject: turbo prolog
Recent reviews have correctly pointed out that turbo prolog's
attempt to enforce type checking has both good and bad points and
that the speed is not very impressive, since much of the
unification can be done at compile time if data types are known.
The major difficulty with Borland's approach to adding strong
typing to prolog is the loss of higher-order predicates. Since a
domain can be at most the disjunction of a small number of
_predeclared_ terms, it is impossible to write a general higher-
order procedure.
This means that you can't write findall, as described in Clocksin
and Mellish (Borland has of course, in their wisdom, provided
such a function). The function doall also cannot be written. It
is handy as a substitute for findall when the predicate Q is
executed for effect only.
doall(Q) :- Q,fail.
doall(_).
First class procedural objects are, in many senses, a much more
fundamental distinction between symbolic and conventional
languages than are heap allocated data structures. Their loss
makes many advanced applications nearly impossible.
------------------------------
Date: 24 Jun 86 11:51:26 GMT
From: decvax!mcnc!duke!jds@ucbvax.berkeley.edu (Joseph D. Sloan)
Subject: Language paradigms
> Can anyone supply me with pointers to readable introductions
> to access-oriented programming? How about articles or
> books on programming paradigms in general? Reply by mail
> and I will summarize results if there is enough interest.
> Joe Sloan,
> Box 3090
> Duke University Medical Center
> Durham, NC 27710
> (919) 684-3754
> duke!jds,
As promised, a highly edited summary follows. Many thanks to
all who replied.
_______________________________________________________________________________
You probably want to find out about a programming system called LOOPS
which was made at PARC in 1981. It combines Procedure-Oriented (like
Lisp) with Object Oriented (like Smalltalk) with Access Oriented (a
program monitors another and gets triggered when a value changes (good
debuggers have watchpoints)), and Rule-oriented (like production/expert
systems).
Bobrow, et al., The LOOPS manual. Tech Rep. KB-VLSI-81-13, Knowledge
Systems Area, Xerox Palo Alto Research Center.
_______________________________________________________________________________
There is a special issue of IEEE SOFTWARE (Jan '86) on "multiparadigm
languages and environments" which may be of some help to you.
_______________________________________________________________________________
AA programming's also mentioned briefly in "Knowledge Programming in LOOPS:
Report on an Experimental Course", by Stefik, Bobrow, Mittal, and
Conway, in AI Magazine, Fall 1983.
_______________________________________________________________________________
Bobrow, D. G. and Stefik, M. "Perspectives on AI Programming", Science
Feb. 28, 1986
Stefik, M. Bobrow, D. and Kahn, K., "Integration of Access Oriented
Programming in a Multiparadigm Environment", IEEE Software, January 1986
Stefik, M. and Bobrow, D. G. "Object Oriented Programming, Themes and
Variatations" AAAI Magazine, Winter 1986
_______________________________________________________________________________
You might like to chase up the work of Kristen Nygaard if you are not
already familar with it. As one of the designers of Simula, he can
reasonably be said to have invented the whole idea of Object Oriented
Programming - about 30 years ago! I suggest you follow up references
in 10th ACM POPL and 11th Simula-67 Users' conference. Also
Sigplan 20.6. There's also a paper in "Integrated Interactive
Computing Systems" Delgano & Sandewall (Eds), North Holland 1983.
_______________________________________________________________________________
Very worthwhile reading and examples can be found in:
The Structure and Interpretation of Computer Programs
Abelson & Sussman
MIT Press, 1985
A couple of watershed papers are:
Control Structures as Patterns of Passing Messages
Carl Hewitt
Journal of AI, V8 #3, (also, I believe, in: AI, a MIT Perspective)
Definitional Interpreters for Higher Order Programming Languages
John Renolds
Proc. ACM Annual Conf. Aug '72
Reflection and Semantics in Lisp
Brian Smith
ACM POPL 11, 1984
_______________________________________________________________________________
An excellent book on the structure and superstructure of programming is
``A practical handbook for software development'' by N.D.Birrell and
M.A.Ould, Cambridge University Press, 1985. The book is based around
the dataprocessing environment, but can, and should be, applied outside
that area.
------------------------------
Date: Wed, 25 Jun 86 09:26 EDT
From: Seth Steinberg <sas@BBN-VAX.ARPA>
Subject: Re: Doing AI Backwards
Yes, memory seems to be a scarce resource. There was an article in
Science on learning in bees, which explains that bees tend to collect
pollen from one type of flower during a period of time because there is
a cost to learning about a new one. In addition, learning a new flower
squeezes out knowledge about other previously learned flowers.
In other words, a bee can be an expert on one kind of flower at a time
because of memory limitations.
There have been a number of interesting bee articles lately. Writing a
computer system to emulate a bee's behavior might be an interesting
approach. Apparently they can recognize landmarks, learn approaches to
flowers, learn which flowers are obnoxious, communicate locations of
pollen, reason about locations and a host of other things, all in a
brain comparable in size to a large IC.
Seth
P.S. Oh yeah, read the next message. That's right ....
------------------------------
Date: 25 Jun 86 08:04 PDT
From: Newman.pasa@Xerox.COM
Subject: Re: Creativity & Analogy
Take a look at the chapters on this topic in Douglas Hofstadters book
"Metamagical Themas" for an interesting discussion.
>>Dave
------------------------------
Date: Wed 25 Jun 86 13:02:17-PDT
From: Pat Hayes <PHayes@SRI-KL>
Subject: Re: AIList Digest V4 #157
Brian Gaines and I were once both faculty in the same University, and he
explained an interesting and effective technique of leadership called
following from the front. It works like this: suppose one is with a group
of people in a strange place, but someone in the group knows the area: and
its time to go somewhere ( say, to lunch ). Then set off confidently in
some direction or other as though leading the group to the right place. They
will follow you. If its the right way, no problem. If its the wrong way, the
person who knows the right way will say something about how he thinks
the right way is over there..at which point you say something like " h yes,
of course!" and go in the right direction. With a little intelligence
applied to the initial guess, and some practice at conversational bluffing,
this can be quite effective. The end result is that you learn the layout
of the strange area and everyone else in the group thinks of you as someone
worth following. I've seen Brian do this, and it works. Of course, it works
best in areas which have little internal structure and where anyone with a bit
of common sense and a gift with words can come up with something which sounds
like a good direction to move in, and where nobody knows the right way anyway.
Pat Hayes
[There is a related "psychic" technique called muscle reading. The
psychic leaves the room and some object is selected. The psychic returns,
grabs someone's arm, and begins leading him rapidly around the room.
Soon they arrive at the selected object and the psychic identifies it.
The trick, which is reportedly easy to learn, is that the subject being
led provides inertial clues due to his anticipation of search path.
Belief in the psychic's ability may help, but rapid motion is sufficient
to produce reflexive muscle responses. -- KIL]
------------------------------
Date: Wed 25 Jun 86 11:40:31-PDT
From: Pat Hayes <PHayes@SRI-KL>
Subject: Policy - Covert Ads
I know I've flamed about this before and been answered at length, but Matt
Hefrons "query" irritated me. I haven't seen such a good advertisement
masquerading as something innocent since watching Masterpiece Theatre. Matt
wants to survey marketed expert systems: fine. Is it really necessary to tell
us that SxxxPxxx is such a one ( NOT, he is careful to point out to us, a
mere shell ), marketed by some company ( whose name he is careful to spell out
for us ), and which ( just in passing we can infer ) runs on - gosh - a PC.
The query could have been stated quite clearly without all this commercial
hype spraypainted over it.
Pat Hayes
------------------------------
End of AIList Digest
********************
