@SRI-AI.ARPA:LAWS@SRI-AI.ARPA (03/30/85)
From: AIList Moderator Kenneth Laws <AIList-REQUEST@SRI-AI>
AIList Digest Saturday, 30 Mar 1985 Volume 3 : Issue 42
Today's Topics:
Administrivia - AIList Mailboxes,
Applications - Chemistry AI Expert Systems & Spelling Check Algorithms &
Planning in a Dynamic Environment & Associative Processing,
Help Wanted - AI Lecturer,
AI Tools - MacIntosh Lisp,
Games - GO,
Recent Articles - Expert System Shells & Survey,
Linguistics - Development of Pidgin, Creole, and NL,
Description - Edinburgh Intelligent Knowledge Based Designer System,
Conference - Workshop on Expert Systems
----------------------------------------------------------------------
Date: Fri 29 Mar 85 09:51:10-PST
From: Ken Laws <Laws@SRI-AI.ARPA>
Reply-to: AIList-Request@SRI-AI.ARPA
Subject: AIList Mailboxes
I have recently been getting many messages in inappropriate mailboxes.
In most cases I can deduce the sender's intention and forward to the
appropriate mailbox, but it would be a help to me if readers would
observe the following convention:
AIList@SRI-AI.ARPA - Submissions for broadcast to the list.
AIList-Request@SRI-AI.ARPA - Messages related to list administration
or policy, for private reply.
Laws@SRI-AI.ARPA - Messages unrelated to AIList policy.
-- Ken Laws
------------------------------
Date: Thu, 28 Mar 85 12:38:12 EST
From: Morton A Hirschberg <mort@BRL-BMD.ARPA>
Subject: Request for chemistry AI expert systems
I am looking for information about expert systems in chemistry. More
specifically, documentation if any exists, such as rules or flow
charts. They need not be in the public domain. Of course, more
information is better (prices for commercial stuff). Thanks.
Mort
[Some of the famous systems are Stanford's DENDRAL and META-DENDRAL
for mass spectrometry and NMR analysis, SUNY(Stonybrook)'s SYNCHEM
system for chemical synthesis, Stanford's MOLGEN and GA1 for DNA
analysis and synthesis, Stanford's CRYSALIS for protein
crystallography, UCSC's SECHS for chemical synthesis, and Rand's
SPILLS for locating and identifying chemical spills. I don't know
which of these are commercial systems, though IntelliCorp (formerly
IntelliGenetics) has derived commercial systems from some of them.
For rules and other info I would suggest that you do a search of the
chemical literature or contact the universities for technical reports;
the AI conference proceedings and journals would discuss mostly the
data structures and reasoning methods. -- KIL]
------------------------------
Date: 03/28/85 13:34:17
From: ADIS@MIT-MC
Subject: Spelling Check Algorithms
[Forwarded from the MIT bboard by SASW@MIT-MC.]
Does anyone know of some good references for spelling checking
algorithms? Human or written references acceptable.
Andy diSessa (ADIS@MC)
[Check the following from Communications of the ACM:
J.L. Peterson, Computer Programs for Detecting and Correcting
Spelling Errors, Vol. 23, No. 12, Dec. 1980, pp. 676-687.
Replys in Vol. 24, No. 5, May 1981, pp. 322 and 331-332;
Vol. 24, No. 9, Sep. 1981, pp. 608-609; and Vol. 25, No. 3,
Mar. 1982, pp. 220-221.
P. Robinson and D. Singer, Another Spelling Correction Program,
Vol. 24, No. 5, May 1981, pp. 296-297, followed by R. Nix,
Experience witha Space Efficient Way to Store a Dictionary,
pp. 297-298. Replys in Vol. 24, No. 9, Sep. 1981, pp. 618-619
[from a pseudonymous Joaquin Miller], and Vol. 25, No. 2,
Feb. 1982, p. 159.
M. Mor and A.S. Fraenkel, A Hash Code Method for Detecting
and Correcting Spelling Errors, Vol. 25, No. 12, Dec. 1982,
pp. 935-938.
D.J. Dodds, Reducing Dictionary Size by Using a Hashing Technique,
Vol. 25, No. 6, June 1982, pp. 368-370.
J.J. Pollock and A. Zamora, Automatic Spelling Correction in
Scientific and Scholarly Text, Vol. 27, No. 4, Apr. 1984, pp. 358-368.
These papers provide references to dozens of others. -- KIL]
------------------------------
Date: Thu, 28 Mar 85 08:57:42 pst
From: coates%usc-liddy%usc-cse.csnet@csnet-relay.arpa
Subject: World models for Planning in a Dynamic Environment
Does anyone know of research on representation scemes for
world knowledge that facilitate planning for an agent which actually
executes in a dynamic domain?
I am interested in modelling an agents ability to detect
anomalies iplan due to unexpected results during plan execution.
Additional complications in plan enactment may occur if the world
contains other agents whose behavior is unpredictable.
If anyone can recommend papers on appropriate knowledge
representations and/or methods for anomally detection in dynamic
worlds contact me at: COATES%USC-liddy@USC-CSE.CSNET
[I have sent a copy of Les Gasser's reference list. -- KIL]
------------------------------
Date: Fri 29 Mar 85 16:14:33-PST
From: MOHAN@USC-ECLC.ARPA
Subject: Associative Processing
I am working on identifying system software requirements for an array processor
(based on associative processing and cellulaer array processing). The processor
is to be used primarily for Artificial Intelligence and Image Undersatnding
tasks. Main system software requirements would be in the area of a suitable
language, its compiler and an operating system. (A host computer is assumed
to be attatched to this processor).
Some pointers to relevant work and literature will be welcome. Please send
mail to me or to AIList.
Thanks.
Rakesh Mohan
ARPA- mohan@eclc
US Mail- Rm #224 Powell Hall
University of Southern California
Los Angeles, CA 90007.
------------------------------
Date: Thu, 28 Mar 85 21:23:43 pst
From: jeff@aids-unix (Jeff Dean)
Subject: wanted: travelling AI lecturer
I've just received a request from someone at California State
Polytechnic (in San Luis Obispo) for a lecture on "AI" (yes, that
provides considerable leeway). Unfortunately, San Luis is a
little out of the way for most of us, being halfway between SF and
LA (a four hour drive from either place). However, if there are
any ("qualified") folks out there who might be interested in
making a presentation, please let me know. Thanks...
P.S. There is no expiration date on this opportunity, but
the presentation should be given during the school year.
------------------------------
Date: Fri 29 Mar 85 09:18:40-PST
From: Ken Laws <Laws@SRI-AI.ARPA>
Subject: MacIntosh Lisp
I have heard a rumor that Expertelligence of Santa Barbara has now
come out with their version of a Maclisp/Commonlisp for the 512K
MacIntosh, priced just below $500. I haven't checked it out, but
(805) 969-7871 was given as the phone number for more information.
-- Ken Laws
------------------------------
Date: 29 Mar 1985 0555-PST
From: MEYERS%UCI-20A@UCI-ICSA
Subject: game of GO
In response to a query about Go programs:
Wilcox, Bruce and Walter Reitman
The Structure and Performance of the Interim.2 Go Program.
IJCAI, 1979. pp.711-719.
Address: University of Michigan
205 Washtenaw Place
Ann Arbor, Michigan 48109
Unfortunately, this is the most recent work I know of; it references
most other work. Theirs is a long-term, conceptual approach. Also:
David J H Brown
Hierarchical Reasoning in the Game of Go.
IJCAI, 1979. pp.114-116.
Address: Computer Science Department
Teesside Polytechnic
Middlesborough, Cleveland, England
Good luck!
Amnon Meyers (meyers @ uci) (2-dan)
------------------------------
Date: 29 Mar 1985 09:52-EST
From: leff%smu.csnet@csnet-relay.arpa
Subject: Recent Articles - Expert System Shells
Infoworld April 1, 1985 page 46 Volume 7 no 13
There is a review of a revised version of "Expert Ease." This is a
system which sets up a classification system based upon examples given
by the user. Many feel that it is not a true "AI" system but rather a
decision table development tool that is being marketed by people as an
AI system for the purposes of making a fast buck.
It is has dropped in price from $2,000 to $695.00 and is being marketed
by Human Edge Software. The ratings are:
two out of a possible four diskettes
performance: good
documentation: fair
ease of use: good
error handling: excellent
support: good
____________________________________________________________________________
Electronics Week, March 25, 1985 page 35
NIXDORF has announced an expert system shell which runs on its 32-bit
minicomputer, the 8832. The cost is $47,00 at current exchange rates.
Nixdorf is also selling a system called Twaice which is designed to
build expert systems to help diagnose malfunctioning objects.
There is also a joint venture between the British Racal Electroinics
and Norway's Norsk data selling an expert system shell. Matra S. A. is
selling an AI system for military training.
Also there are a few paragraphs on GCLisp's Golden Common Lisp.
------------------------------
Date: 28 Mar 1985 18:10-EST
From: leff%smu.csnet@csnet-relay.arpa
Subject: Recent Articles - Survey
ComputerWorld March 18, 1985 "Engineers behind Expert Systems"
A reprint from Patel Harmon, Dave King "AI in business"
____________________________________________________________________________
JACM Vol 32 no 1 Jan 1985
1 Three Approaches to Heuristic Search in Networks A. Bagchi and A. Mahanti
28 And/Or Graph Heuristic Search methods
____________________________________________________________________________
IEEE Trans on Industrial Electronics Vol 32 No 1 Feb 85
Design and Implementation of a Binocular-Vision System for Locating
Footholds on a Multi-Legged Walking Robot F. Ozguner S. J. Tsai Page 26
____________________________________________________________________________
ComputerWorld March 25 1985 Page 11
Usefulness of micro expert systems called limited
------------------------------
Date: Wed, 27 Mar 85 18:40 CDT
From: Patrick_Duff <pduff%ti-eg.csnet@csnet-relay.arpa>
Subject: Development of Nat. Langs., Pidgin, Creole, Lang. Trans/Unders/Gener.
I have a newspaper clipping which quotes Dr. Laurence McNamee, a
linquistics professor at East Texas State University, as follows:
"Theories on the origin of language became so rife and
so romantic that in 1886 the Linguistic Society of Paris
passed a resolution outlawing any more theories on this
topic, a resolution that has since been reaffirmed.
"The truth is that scholars simply do not know how
languages developed because they have never observed one
develop. We have seen many languages die, and recently
even witnessed the rebirth of a language (Hebrew), but
never a language from its initial stage."
Can anyone elaborate on the resolutions and why the issue came up?
Derek Bickerton writes in "Creole Languages" (Scientific American,
July 1983, pages 116-122) that new languages have developed...
"... many times over the past 500 years among the
children of slaves and laborers who were pressed into
service by the European colonial powers.
"These laborers, who were shipped from many parts
of the world to tend and harvest crops in Africa, the
Indian Ocean region, the Orient, the Caribbean and
Hawaii, were obliged to communicate within their
polygot community by means of the rudimentary speech
system called pidgin. Pidgin speech is extremely
impoverished in syntax and vocabulary, but for the
children born into the colonial community it was the
only common language available. From these modest
beginnings, new native languages evolved among the
children, which are generically called creole
languages. It can be shown that they exhibit the
complexity, nuance and expressive power universally
found in the more established languages of the world."
"... scholars have noted a remarkable similarity
of structure among all the creole languages. It can
now be demonstrated, by considering the origin of
creole language in Hawaii, that similarities among
creoles cannot be accounted for by contact with other
languages, either indigenous or imported. The finding
suggests that what is common to creole languages may
indeed form the basis of the acquisition of language
by children everywhere. There is now an impressive
body of evidence to support this hypothesis: between
the ages of two and four the child born into a
community of linguistically competent adults speaks
a variety of language whose structure bears a deep
resemblance to the structure of creole languages...."
Besides recommending that interested parties read this Scientific
American article, I'm curious about whether it would be useful to use
a creole language as an intermediate language for translation, what
"pidgin speech" with its "impoverished syntax and vocabulary" could
tell us about how to design command languages for computer systems,
whether the current state-of-the-art in natural language parsers and
generators are up to the task of using pidgin or creole, etc.. After
looking at some of the sample sentences in the article, it seems to
me that it would be easier for a computer program to generate or
understand a sentence in Pidgin or Hawaiian Creole than it would be
for it to handle the English equivalent.
regards, Patrick
Patrick S. Duff, ***CR 5621*** pduff.ti-eg@csnet-relay
5049 Walker Dr. #91103 214/480-1905 (work)
The Colony, TX 75056-1120 214/370-5363 (home)
(a suburb of Dallas, TX)
------------------------------
Date: Friday, 29-Mar-85 17:06:35-GMT
From: GIDEON FH (on ERCC DEC-10) <G.Sahar%edxa@ucl-cs.arpa>
Subject: Edinburgh Intelligent Knowledge Based Designer System
UNIVERSITY OF EDINBURGH
DEPARTMENT OF ARTIFICIAL INTELLIGENCE
Alvey Large Scale Demonstrator Project
"Design to Product" at the University of Edinburgh:
The Alvey Directorate have awarded a major contract to a consortium of
companies and universities: GEC Electrical Projects, GEC Marconi Research
Centre, GEC Avionics, Lucas CAV, the National Engineering Laboratory and
Edinburgh, Leeds and Loughborough Universities. The part of the project
to be carried out at Edinburgh will involve the development of a novel
Intelligent Knowledge Based Designer System. This Designer System will
enable a design engineer to communicate interactively the conceptual
function and form of a design, and to interface the resulting product
description to a manufacturing capability. It will be implemented in
POP-11, Prolog and Lisp, running in the Poplog environment, under the UNIX
operating system.
Further information is also to be had from tims%edxa@ucl-cs.arpa.
If you are interested and have questions, don't hesitate to ask them.
------------------------------
Date: Fri, 29 Mar 85 18:03:46 cst
From: porter@anl-mcs (Porter)
Subject: Workshop on Expert Systems
Workshop on Knowledge Engineering/Expert Systems
The twenty-fourth annual workshop sponsored by the Western
Committee of the IEEE Computer Society will be held September 4-
6, 1985 at the UCLA Conference Center at Lake Arrowhead. The
subject of the workshop is: "Knowledge Engineering: How?"
Sessions are planned on knowledge acquisition, knowledge
representation, inferencing strategies, and programming
environments.
Topics of discussion include the following: Are there domain-
specific approaches to knowledge acquisition? How can an expert
system tell when it is in an area outside of its competence?
What is the best way to choose inference strategies? How much
can expert system builders help? How important are user models?
How does one deal with the uniqueness of an expert's knowledge?
How do questions of acquisition relate to representation and
inference strategies?
Due to the limited facilities, attendence will be by invitation.
People working in the knowledge engineering and expert systems
area are encouraged to contact the program chairperson, Greg
Kearsley, Courseware, Inc., 10075 Carroll Canyon Road, San Diego,
Ca 92131, (619) 578-1700 or the general chairperson, Sig Porter,
Merdan Group, Inc. 4617 Ruffner Street, Box 17098, San Diego, CA
92117 (619-571-8565).
(note: Greg will be out of communication until about April 20,
and Sig will also be unreachable until April 8.)
------------------------------
End of AIList Digest
********************