Copy Link
Add to Bookmark
Report
AIList Digest Volume 4 Issue 127
AIList Digest Wednesday, 21 May 1986 Volume 4 : Issue 127
Today's Topics:
Queries - Expert Systems and Engineering Design & TMS Variables,
Games & Literature - Conway's Game of LIFE,
AI Tools - VAX VMS LISP & LISP Utilities &
AI, Graphics, and Simulation,
Expert Systems - ADL Personal Financial Planner,
Perception - Correction to Response to Nevin,
Literature - Research Indexes: Machine Learning 3,
Discussion Lists - Psychnet
----------------------------------------------------------------------
Date: 20 May 86 17:43:41 GMT
From: ernie.Berkeley.EDU!mazlack@ucbvax.berkeley.edu (Lawrence J. Mazlack)
Subject: Re: ES and ENGINEERING DESIGN
I am in the definition stage of my dissertation. I need help in identifying
what is being done and what has done. So, any pointers or information
would be greatly appreciated.
My general area is: ENGINEERING DESIGN AND EXPERT SYSTEMS
Within that, I am interested in: ANALOG CIRCUIT DESIGN
Thank you,
Mojgan Samadar pelejhn@ucccvm1.bitnet 508 Riddle Road, #46
Cincinnati, Ohio 45220
------------------------------
Date: Tue, 20 May 86 16:18:54 CDT
From: mklein@b.CS.UIUC.EDU (Mark Klein)
Subject: TMS Variables
I have been trying to decide whether or not to change some code I have (a
pattern-directed inference engine, justification-, logic-, and assumption-
based TMS's) to handle variables in the data. Variables do seem to add
more flexibility (specifically, universal and existential quantification, I
guess), but have some problems as well - for example, one can no longer
use open-coded unification to match up assertions to rule triggers (I think).
What I'd like to know is whether it is easy or hard to have variables
in data for each kind of TMS (it seems to be simple for
the justification-based TMS)? Is there any advantage to having variables in
data that you can't get simply by writing the appropriate rules (i.e. in
terms of efficiency, expressiveness, ability to control inference, etc)?
Thanks,
Mark Klein
------------------------------
Date: 20 May 86 18:43:12 EDT
From: kyle.wbst@Xerox.COM
Subject: re:Conway's game of LIFE
In AIList Digest V.4 #125, Steve Marsh requested info on any book ("not
too theoretical") on this subject.
I recommend The Recursive Universe, by William Poundstone (Morrow &
Co.), c.1985. It is a delightful non-mathematical treatment of
information theory, entropy relationships and other related issues
including a discussion of Conway's game of LIFE. It also includes a
listing of the game for use on an IBM PC (or equivalent).
Earle Kyle.
------------------------------
Date: Tue, 20 May 86 09:44:24 EDT
From: jbs@EDDIE.MIT.EDU (Jeff Siegal)
Reply-to: jbs@mit-eddie.UUCP (Jeff Siegal)
Subject: Re: VAX VMS LisP
In article <8605200626.AA27708@ucbvax.Berkeley.EDU> larry@JPL-VLSI.ARPA writes:
>Are there any Common LisPs for the VAX under VMS? (DEC's VAX LisP is an
>Ultrix product only, so far as I know.) ...
You have things backwards. VAX LISP is a VMS-only product.
Jeff Siegal
------------------------------
Date: Wed, 14 May 86 21:21:07 edt
From: beer%case.csnet@CSNET-RELAY.ARPA
Subject: VAX LISP Utilities
Here at the Center for Automation and Intelligent Systems Research (Case
Western Reserve University), we have developed a number of tools and utilities
for VAX LISP. They include extensions to the control structure and string
primitives, a simple pattern matcher, a pattern-based APROPOS facility, a
pattern-based top-level history mechanism, an extensible top-level driver,
an extensible top-level command facility, an extensible DESCRIBE facility,
and an implementation of Flavors. These facilities are described in detail
in a technical report, "CAISR VAX LISP Tools and Utilites" (TR-106-86).
The object code for these facilities is in the public domain. A tape
containing them may be requested by sending me mail at the address below.
There is a $35 charge to cover the cost of the tape and shipping and handling
costs. Currently, these facilities assume a VMS operating system environment
and require Version 2.0 of VAX LISP.
Randall D. Beer
(beer%case@CSNet-Relay.ARPA)
Center for Automation and Intelligent Systems Research
Case Western Reserve University
Glennan Bldg., Room 312
Cleveland, OH 44106
------------------------------
Date: Thu, 15 May 86 10:19 EDT
From: Paul Fishwick <Fishwick%upenn.csnet@CSNET-RELAY.ARPA>
Subject: AI, Graphics & Simulation
>> Date: Mon, 12 May 86 12:19:46 est
>> From: munnari!csadfa.cs.adfa.oz!gyp@seismo.CSS.GOV (Patrick Tang)
>> Subject: Graphics, Artifical Intelligence and Simulation
>>
>> Is there anyone out there come across any literature
>> describing the topics Graphics, Artifical Intelligence and
>> Simulation together. It seems to me that literatures on
>> these combined topics are VERY VERY scarce!!!
>>
There are a number of projects that have incorporated graphics, ai, and
simulation. Perhaps the largest project has been the STEAMER project which
incorporates layers of object abstractions for a steam plant. At Penn, we have
a multi-level simulation system (HIRES) that permits the construction and
interactive control of process abstraction layers. We also have a facial
animation system (OASIS) that incorporates local area expression simulations.
Both HIRES and OASIS utilize the Iris Workstation 2400 (Silicon Graphics, Inc.)
for real-time animation. Rand Corporation (Santa Monica, CA) has been involved
with object oriented simulation for quite some time (I think they have an
example graphical simulation of a battle scenario). You should also check out
the commercial enterprise, Pritsker Associates, who sell a graphical simulation
package. Some references are given:
1. Hollan James, Hutchins Edwin, Weitzman Louis - "STEAMER: An
Interactive Inspectable Simulation-Based Training System", AI
Magazine (Summer 1984).
2. Fishwick, Paul - "Hierarchical Reasoning: Simulating Complex
Processes over Multiple Levels of Abstraction", Ph.D Thesis, Univ.
of Pennsylvania, 1986 (MIS-CS-85-21).
3. Platt, Steve - "A Structural Model of the Human Face", Ph.D Thesis,
Univ. of Pennsylvania, 1985.
4. McArthur, David and Sowizral, Henry - "An Object-Oriented Language
for Constructing Simulations", IJCAI 1981.
An important issue with "AI and Simulation" is determining where the "ai" is in
simulation. The answer to that may best be found in the special workshop in AI
and Simulation to be held at AAAI-86. Even though graphics is not explicitly
mentioned, you should also check out the qualitative reasoning/simulation
literature (de Kleer, Forbus, Kuipers, and others) in past IJCAI/AAAI's. Also
"aggregation" is receiving wider attention these days: look at Goldin & Klahr
(IJCAI '81) and Weld (IJCAI '85).
-paul
CSNET: fishwick@upenn
------------------------------
Date: Mon, 19 May 86 19:08 EDT
From: Tom Martin <TJMartin@MIT-MULTICS.ARPA>
Subject: Correction to Spang-Robinson Summary
A message appeared in the AI List a few weeks ago that summarized the
Spang-Robinson report of April, 1985 -- Volume 2, No. 4
The Arthur D. Little personal financial planner, according to the
summary, "uses databases residing on IBM mainframe." In fact, what we've
done is develop a serial link that connects the Symbolics to an IBM 43xx
running VM/CMS via the IBM 3705 communications controller.
The point worth expanding on is that in this IBM/Symbolics symbiosis,
the IBM is in charge, not the Symbolics. Yes, the planner "uses
databases on the IBM mainframe," but only because the IBM user has (a)
requested it, and (b) has the right access.
The system uses the "virtual machine" concept of VM/CMS heavily. In
effect, the Symbolics appears to be a separate virtual machine running
on the 43xx. It responds to commands in real-time (instead of
uploading/downloading files), however, the commands have to be
pre-defined or rely on the EVAL on the Symbolics to return a meaningful
answer.
Neither the PFPS or this link are products of Arthur D. Little, Inc.,
as such. I hope to get a chance to talk about the IBM link at SLUG in a
few weeks.
Tom Martin Manager -- AI Systems Development
------------------------------
Date: Sat, 17 May 86 15:53:31 PDT
From: kube%cogsci@BERKELEY.EDU (Paul Kube)
Subject: correction to response to Nevin, AIList V4 #120
Instead of
Since the standard deviation of the difference of two
identically normally distributed random variables is twice the
standard deviation of either variable, the temporal disparity
measurement has 95% confidence interval of 4*10^-6 seconds.
I should have said
Since the standard deviation of the difference of two
identically normally distributed random variables is less than
twice the standard deviation of either variable, the
temporal disparity measurement has 95% confidence interval
of less than 4*10^-6 seconds.
The argument for getting to 4*10^-7 second accuracy still holds.
Paul Kube
Berkeley EECS
kube@berkeley.edu
------------------------------
Date: Fri, 16 May 86 13:29:08 EDT
From: Bruce Nevin <bnevin@cch.bbn.com>
Subject: bad enough theory
Response to kube%cogsci@berkeley.edu (Paul Kube)
I suppose we ought to tell the authors of that fish story, and the editors
of Nature, that their experimental results have a trivial interpretation
in elementary signal detection theory. They were all struck by the
`impressive computational ability' of the central nervous system in these
animals. I'll leave it to you to disabuse them of that.
Your account succeeds precisely because it is not reductionist.
Thanks for clarifying what is going on there.
------------------------------
Date: 18 May 86 13:25 PDT
From: Shrager.pa@Xerox.COM
Subject: Research indexes: Machine Learning 3
Conference proceedings are often published as books for which the
authors have not taken the time to construct a proper index. The index
for the most recent machine learning book (ML3), is extremely poor. In
a book containing 77 articles, claiming to be a "guide to current
research", the index is clearly the most important part of the book. It
is a shame that this is really a valuable source of summary articles
because its success will lead the publisher and authors to believe that
they can sell poorly indexed books.
Although ML3 is a summary of current research, it is surely no guide. If
one's research is actually guided by consideration of the index to ML3,
then the authors and publisher have done a disservice to the field.
In this particular case, I am surprised that this book wasn't properly
indexed. The authors were asked to submit SCRIBE source for their
papers and could easily have been asked to also include index entries
either in the source, or selected from a list of keywords.
I encourage those who feel that they must buy ML3 to ignore the index,
and to write a letter to the authors author and publishers complaining
about its index. The 22 cents spent on stamps, and the hour spent
writing, are nominal with respect to the service you may be doing
yourself and the field in the future.
-- Jeff
------------------------------
Date: Mon, 12 May 86 21:58:25 pdt
From: George Cross <cross%wsu.csnet@CSNET-RELAY.ARPA>
Subject: Psychnet
[Ken - I think this may be of interest to some portion of the AI community.
It was on the Weizmann Bulletin Board]
PPPP IIII
PP II
PP PSYCHNET II
PP II
PP II
PP II
PSISSSSPSI
SSSS
SSSS
SSSS
SSSS
SSSS
SSSS
PSISSSSPSI
EDUCATIONAL PSYCHOLOGY DEPARTMENT
UNIVERSITY OF HOUSTON - UNIVERSITY PARK
4800 CALHOUN BOULEVARD
HOUSTON, TEXAS 77004
For More Information:
Norman Kagan, Ph.D. or
Distinguished Professor
Robert Morecock, M.A.
Graduate Assistant
713-749-7621
EPSYNET@UHUPVM1.BITNET
March 21, 1986
Selected professional papers scheduled for presentation at the
August APA convention soon will be available worldwide via
PSYCHNET, the new electronic bulletin board and mail machine at
the University of Houston.
Dr. Norman Kagan, distinguished professor and chair of the
Educational Psychology Department there said, "PSYCHNET will
enable psychologists to arrive with papers in hand, having read
and processed the presenters' ideas months in advance of the
convention. They will be ready to discuss these ideas rather than
just assimilate them at our annual meeting."
PSYCHNET, which will eventually serve the psychological
community in a number of ways, is distributed by BITNET, the
electronic mail network that links much of the academic world.
"This year PSYCHNET will be sending out papers for Divisions
12, 16, 17 and 38 to anyone on the BITNET system who requests
them. Next year we hope that even more divisions will
participate," continued Dr. Kagan, "and we are already encouraging
people to contact us via BITNET because PSYCHNET is up and
running."
"Most requests for papers will be automatically acknowledged
within five seconds," said Dr. Kagan. Actual arrival time of
requested papers will vary from five seconds to twenty minutes,
depending on how busy the mail system is. "We are hoping people
will go ahead and give PSYCHNET a try now, rather than wait till
the last minute."
For most computer sites with VM operating systems the command
TELL UH-INFO AT UHUPVM1 PSYCHNET HELP will start the process of
requesting PSYCHNET papers. Many VAX sites (JNET) will find the
command
SEND UH-INFO@UHUPVM1 PSYCHNET HELP
will obtain the same information. Others should consult
their local computer center regarding sending the message PSYCHNET
HELP to userid UH-INFO at node UHUPVM1.
Via leased lines BITNET makes PSYCHNET available at some 1203
nodes or university computer sites in the free world. Locations
range from Israel in the Middle East, to Europe, the United States
and Canada, then west as far as Tokyo in Japan.
Users of VM (BITNET) systems can obtain the psychnet
exec, which vastly simplifies using psychnet. Just give the command
TELL UH-INFO AT UHUPVM1 PSYCHNET SENDME PSYCHNET EXEC
When it arrives move it from your reader to your file area and then
give the command PSYCHNET. Your PSYCHNET use will now be automatic
------------------------------
End of AIList Digest
********************