Copy Link
Add to Bookmark
Report
AIList Digest Volume 4 Issue 214
AIList Digest Tuesday, 14 Oct 1986 Volume 4 : Issue 214
Today's Topics:
Query - Connectionist References,
Applications - Animal Rule-Systems Simulations,
Expert Systems - Coupling Numeric and Symbolic Computing,
Survey - Interactive 2-D Math Editing Interfaces
----------------------------------------------------------------------
Date: Tue, 14 Oct 86 14:48:20 nzs
From: ubc-vision!calgary!vuw90x!paul@seismo.CSS.GOV (Paul Fitchett)
Reply-to: ubc-vision!vuwcomp!paul@seismo.CSS.GOV (Paul Fitchett)
Subject: Connectionist References Request
Recent items in mod.ai have piqued my interest about "connectionist"
ideas in AI. I wonder if anyone could provide a number of references
that are a good introduction to the ideas in this area.
The little I've read makes them seem like perceptrons -- I hope not.
If replying by email please use one of the paths below.
Thanks,
Paul
Paul Fitchett
uucp : ...!{ubc-vision, alberta}!calgary!vuwcomp!paul
ACSnet : paul@vuwcomp.nz
subethernet : ...local-group!milky-way!sol!terra!nz!vuwcomp!paul :-)
------------------------------
Date: 13 Oct 86 18:14:19 GMT
From: ucdavis!deneb!g451252772ea@ucbvax.Berkeley.EDU (g451252772ea)
Subject: animal rule-systems simulations
By way of introduction to the following Mail message, 'bc' posted last
spring a query for anyone with references on 'simulation animal behavior using
rule-driven systems'. I discovered his message in an old listing, and find
the topic of interest also. In case 'bc' (William Coderre) is no longer
at mit-amt.MIT.EDU, I'm posting to the net also. Thanks for your tolerance...
Hi, bc (?bc?):
I'm curious about any replies you got to your query last April for
rule-driven simulations of animal behaviors. I have somewhat
similar interests, reflecting my grad work in ethology here at Davis, and my
undergrad work at U.C. Santa Cruz in Information- Computer science. We have
here a person doing stuff you'd enjoy: Marc Mangel, with his 'dynamic
stochastic optimization' analysis of everything from insect oviposition
choices to foraging theory to fisheries harvesting. His insight seems to be
the addition of a 'state variable' - usually characterized as energy
reserves, gut contents or similar - to revamp the static optimization models
of Houston, McNamara, Krebs, et al. (the 'Oxford' crowd). Mangel is chair
of the math dept. here, and co-authors with Colin Clark of the U. British
Columbia. Clark is visiting here this quarter and giving an applied math
seminar, with lots of application studies. Both guys emphasize computer
programs, and the programs have a game-like air to them. If you'd like more
info, I can send some typed notes by Mangel describing the analysis, and one
of his most counter-intuitive applications.
Mostly I'm working with evolutionary studies: the predator/prey
interactions of snakes and ground squirrels (my thesis is on stupidity: the
dumbness of Arctic ground squirrels, which don't even appear to <recognize>
snakes of any kind, much less handle them correctly). I do have to give a
week's worth of lectures to my animal-behavior group next month, explaining
'artificial intelligence' ab initio to them. Despite Mangel and Clark, the
prejudice against math/systems here is substantial. Any ideas you have for
good material/examples, in the vein of Winograd's new book or Rosen's
discussion of ANTICIPATORY SYSTEMS (much watered down!) or ANYTHING else,
would be most welcome!
Thanks --Ron.
------------------------------
Date: Fri 10 Oct 86 15:00:40-EDT
From: Albert Boulanger <ABOULANGER@G.BBN.COM>
Subject: Expert Systems & Math Models
Check out the book:
Coupling Symbolic and Numerical Computing in Expert Systems
Edited by J.S. Kowalik and is based on the workshop by this name held
at Bellevue, Washington 27-29 August, 1985.
Elsevier, 1986
------------------------------
Date: 0 0 00:00:00 PDT
From: "LLLASD::GARBARINI" <garbarini%lllasd.DECNET@lll-crg.arpa>
Reply-to: "LLLASD::GARBARINI" <garbarini@lllasd.decnet>
Subject: RE: Availability of interactive 2-d math editing
interfaces...
The following is a summary of responses to my query on 2-d math editing
interfaces.
I'd like to thank everyone who responded. I hope to eventually respond
to each of you individually.
----------
Joe P. Garbarini Jr.
Lawrence Livermore National Lab
P. O. Box 808 , L-308
7000 East Avenue
Livermore Ca. , 94550
arpanet address: GARBARINI%LLLASD.DECNET@LLL-ICDC.ARPA
----------
The original query:
-----
I am working with a number of other people on a project called Automatic
Programming for Physics. The goal is to build an AI based automatic
programming system to aid scientist in the building of numerical
simulations of physical systems.
In the user interface to the system we would like to have interactive
editing of mathematical expressions in two-dimensional form.
It seems a number of people have recently made much progress in this
area. (See C. Smith and N. Soiffer, "MathScribe: A User Interface for
Computer Algebra Systems," Conference Proceedings of Symsac 86, (July,
1986) and B. Leong, "Iris: Design of a User Interface Program for
Symbolic Algebra," Proc. 1986 ACM-SIGSAM Symposium on Symbolic and
Algebraic Manipulation, July 1986.)
Not wishing to reinvent the wheel, I'd appreciate receiving information
regarding the availability of any such interface.
=======================================================================
From: Joe Garbarini (Yes, this is from me!)
MathSoft makes a product call MathCAD which has an interactive 2-d math
interface. Currently runs on IBM PCs.
Mathsoft, Inc.
One Kendal Square, Bldg. 100
Cambridge, MA 02139
800-628-4223
-----
From: James E. O'Dell <jim@ACG.arpa>
Normal MACSYMA has 2-d editing done by Carl Hoffman and Rich Zippel.
I think references to it can be found in one or the other of the
Proceedings of the MACSYMA Users Group.
Jim
-----
From: fateman@dali.Berkeley.EDU (Richard Fateman)
There is some stuff on Sun-2 equipment working with macsyma
here at UC Berkeley. The MathScribe stuff is currently nicer
looking in my opinion, but people are still working on stuff
here.
<and from elsewhere this:>
-----
A version of Macsyma for the VAX computer, including sources and binaries
for Macsyma and the underlying Lisp (Franz Lisp opus 38.91), is in the
National Energy Software Center library. (Argonne, IL.)
(312) 972-7172. This should run without change on 4.3BSD UNIX or ULTRIX.
This version, which was developed at the University of California, has
also been run, with modifications, on various other (non-VAX) systems
which support the Franz Lisp dialect. For information on Franz Lisp
for VAX/VMS or other computers, you might wish to contact your hardware
vendor or Franz Inc. in Alameda CA, (415) 769-5656. (for many mainframe
and workstation computers)
Vaxima uses about 4.5 megabytes of address space to start up, and
as configured, can grow to 6.5 megabytes or so. By changing a compile-time
parameter in the Lisp system, the system may be configured to grow much
larger. (We have run a 53 megabyte system on a VAX 8600).
At UC Berkeley we have been using this code on Sun-2 and Sun-3 systems,
and microVAX-II's.
Vaxima is quite fast when given enough physical memory, and appears at this time
to be very cost-effective compared to implementations on special-purpose
Lisp machines or "DOE-MACSYMA" for VAX/VMS.
There are a number of packages that have been developed to work with this
program (e.g. user interfaces, better algorithms for factoring, graphics,
an interface to the Numerical Algorithms Group (NAG) library).
Neither UC Berkeley nor by NESC provides support for vaxima.
Richard J. Fateman, U.C. Berkeley
-----
From: arnon.pa@xerox.com
For some years Xerox has offered software for interactive
two-dimensional editing of mathematical expressions as part of its Star,
and now Viewpoint, systems. Regrettably Viewpoint runs only on Xerox
workstations.
As part of the research programming environment at Xerox PARC, we have a
more powerful math editing and display package which is roughly the
equivalent of MathScribe. A noteworthy property of our package is that
a math expression can be moved interchangeably between the editor, a
technical document, and a system for symbolic mathematical computation.
I'd be happy to discuss or demo.
Dennis Arnon
Computer Science Laboratory
Xerox Palo Alto Research Center
3333 Coyote Hill Road
Palo Alto CA 94304
(415) 494-4425
-----
From: seismo!Xerox.COM!Kelley.pa
While not oriented specifically to physical systems, STELLA for the
Macintosh from High-Performance Systems, Inc., 13 Dartmouth College
Highway, Lyme, New Hampshire 03768. has a very nice user interface for
help in building mathematical models. It is oriented toward Forrester's
Systems Dynamics simulation paradigm. The user interface is worth
examining.
-- kirk
-----
From: mcgeer%sirius.Berkeley.EDU@BERKELEY.EDU (Rick McGeer)
soiffer@dali.berkeley.edu
carolyn@dali.berkeley.edu
are the addresses of Neil Soiffer and Carolyn Smith, respectively. Benton
Leong is at
allegra!watmath!blleong
<Rick, thanks for the info for the rest of the project ---JPG>
-- Rick
-----
From: Doug A. Young <dayoung%hplabsc@hplabs.HP.COM>
Tony Hearn forwarded a message from you a while back to me, asking
about an interface for algebra systems. I did my masters thesis on
a graphical multi-window system for Macsyma. If you are interested,
I could forward some information on it to you. Contact me at
dayoung@hplabs
Doug Young
-----
From: Bill Schelter <ATP.SCHELTER@R20.UTEXAS.EDU>
I have an emacs like editor which allows the display
of mathematics as well as textual material.
You can mouse into the superscript position etc.
The system is called INFOR, and is available for lisp
machines (ymbolics and TI
I also ported the version of macsyma from doe, to run
on those machines. It would be fairly easy to
connect the two, since they are running in the same memory
space.
The display is very good, at least as good as TEX.
You can actually create a dvi file directly from the editor.
This can be then printed to obtain very high quality output.
Bill Schelter
-------
From: mcvax!fransh@seismo.CSS.GOV (Frans Heeman)
Some while ago, you put a query on the news about the availability
of interactive 2-d math editing interfaces. We are working on
a formula-editor (NOT a formula-manipulator). The idea is as
follows:
For example, the user gives the command for a fraction
to be entered. On the screen a small horizontal bar
is displayed, and the cursor is positioned above
this bar. The user types in the nominator.
While typing, the fraction-bar remains as long as the
nominator. Next the user gives the END-command, to indicate
the end of the nominator. Now the cursor is centered under
the fraction-bar, and the user types in the denominator. While
typing, the nominator and denominator remain centered with
respect to the fraction-bar, and the fraction-bar remains as
long as the longest of the nominator and denominator.
By means of keyboard and mouse (menu's) the user can enter a
mathematical formula. While typing, the formula is at every
moment displayed on the screen in its current 2-dimensional
form. The system can handle mathematical constructs as fraction,
root, integral, matrices, etc. The system also handles greek and
italic characters. The constructs the system can handle are specified
in an external grammar, so it is relatively easy to add or change
constructs. The way the formula is displayed on the screen is
also specified in this grammar. It is possible to get a
hard-copy of the formula: this is done by generating 'eqn'-code
for the formula, and then get a typeset result by using
'eqn' and 'troff' (part of the UNIX operating system). Finally, a
formula can be saved, retrieved and edited.
Our system is still under development and not yet available
for 'real' use. The final goal is to make a document-editor for
text, tables, mathematical formulae and (simple) pictures.
In France, Vincent Quint et al is doing work along these
lines too with a system previously called 'Edimath', currently
called 'Grif'.
References:
We have only published an internal report (in dutch), and are
currently preparing an English article to be published.
Edimath:
V. Quint (March 1983),
"An Interactive System for Mathematical Text
Processing",
Technology and Science of Informatics, vol. 2, nr. 3,
pp. 169-179.
Grif:
V. Quint, I. Vatton (April 1986),
"Grif: An Interactive System for Structured Document
Manipulation",
Proceedings of the Conference on Text Processing and
Document Manipulation,
Nottingham, England, 1986.
Frans C. Heeman
Centre for Mathematics and Computer Science (CWI)
P. O. Box 4079
1009 AB Amsterdam
The Netherlands
fransh@mvcax.UUCP
------------------------------
End of AIList Digest
********************
