Copy Link
Add to Bookmark
Report
AIList Digest Volume 5 Issue 222
AIList Digest Tuesday, 29 Sep 1987 Volume 5 : Issue 222
Today's Topics:
Bindings - Gary Cottrell & Slava Prazdny,
Review - Spang Robinson #3/9,
Announcement - Second Chair in AI at Sussex University,
Neural Networks - Hamming Classification Network
----------------------------------------------------------------------
Date: Wed, 23 Sep 87 21:52:35 PST
From: gary@sdcsvax.ucsd.edu (Gary Cottrell)
Subject: binding
I am now at:
Gary Cottrell
Department of Computer Science and Engineering C-014
UC San Diego
La Jolla, Ca. 92093
gary@sdcsvax.ucsd.edu (ARPA)
{ucbvax,decvax,akgua,dcdwest}!sdcsvax!sdcsvax!gary (USENET)
619-534-6640
------------------------------
Date: Mon 28 Sep 87 19:47:21-PDT
From: Ken Laws <Laws@KL.SRI.Com>
Subject: Slava Prazdny
Date: Wed, 23 Sep 87 15:38:40 PDT
From: Amy Lansky <lansky@venice.ai.sri.com>
Subject: sad news
I thought I should let people know of a very unfortunate accident that
occurred this weekend. Slava Prazdny (formally at SPAR/FLAIR, most recently
at FMC) died this past Sunday in a hang-gliding accident. There will
be a memorial service October 7. I will let you know of more details
when I find out.
-Amy
------------------------------
Date: Mon, 28 Sep 1987 10:56 CST
From: Leff (Southern Methodist University)
<E1AR0002%SMUVM1.BITNET@wiscvm.wisc.edu>
Subject: Summary of Spang Robinson #3/9, 9/87 (Leff file bm788)
Summary of Spang Robinson Report, September 1987, Volume 3, No. 9
The lead article is on neural networks.
TRW's Mark III offeres 500K interconnections per second for an eight board
machine.
Price for the Mark III is $60,000 to $90,000.
Robert Hecht-Nielsen and Todd Gutschow formed HNC which produces
a board level neurocomputer and software costing $9500 to $19,500.
Science Applicatiosn Corporation system does 10M interconnections
per second called the Delta-1 costing $15,000.
Texas Instruments, Siemens, AT&T Bell Labs and Synaptics are developing
true analog neural network chips.
Martingale Research has a contract with Wright Aeronautical Labs to work with
biological networks in culture. It also sels a network simulator software
costing from $75.00 to $1275.00.
Meiko Incorporated has sold 100 of the The Computing Surface which
supports from 1-1024 processor nodes.
Nestor Incorporated sells a data entry system for handwriting input to
computers for $1595, a "Decision Learning System" and the Nestor Development
system. It had ~$400,000 revenue in FY 86.
Neuraltech sells Plato/Aristotle which is a system development kit costing
$2000.00 for "Beta version."
Neuralware sells a neural network prototyping and developing system for $495.00
They have a 200 order backlog.
+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_+_
Shorts
Symanetec is merging with Living Video Text (inventor of Think Tank, Ready
and More).
Borland bought Ansa, the maker of Paradox. Borland has sold 120,000
copies of Turbo Prolog.
Information Builders Inc., of Focus fame, acquired Level Research.
Intellicorp reported a net loss of about four million on revenues
of about twenty million for the fiscal year ending June 30, 1987.
Symbolics reported losses of about twenty-five million on revenues of
about 103 million. They have announced version four of their Prolog
system.
Integrated Inference Machines has placed one of its SM45000 symbolic
processing systems at NASA-Ames.
DEC has established an AI Lab at Palo Alto.
Natural Language Incorporated has ported Data Talker and the NLI connector
to Apollo work stations.
Texas Instruments has introduced its Explorer II Color System.
Advanced Decision Systems has been awarded a research
contractor to monitor seismic events resulting from nuclear tests.
James McGowan is now the president and CEO of Palladian Software.
GENSYM of Cambridge, MA has filed a lawsuit against GigaMos
systems, both of whom are offsprings of LMI. GENSYM claims
that GigaMOS is "interfering with Gensym's customers and prospects."
------------------------------
Date: Wed, 23 Sep 87 09:18:12 gmt
From: Aaron Sloman <aarons%cvaxa.sussex.ac.uk@NSS.Cs.Ucl.AC.UK>
Subject: Second Chair in AI at Sussex University
I'd be grateful if you are able to post this. Thanks.
Aaron Sloman
UNIVERSITY OF SUSSEX
S C H O O L O F C O G N I T I V E S C I E N C E S
CHAIR IN ARTIFICIAL INTELLIGENCE
================================
Sussex University, a major UK centre for research and teaching in AI,
intends to appoint a second Professor of Artificial Intelligence in the
newly established multidisciplinary School of Cognitive Sciences, which
includes AI, Linguistics, Philosophy and Psychology. Applications will
be welcomed from candidates with research interests in areas of AI
relevant to cognitive processes in natural or artificial systems.
The appointee will play a major role in the continued development of AI
and Computer Science within the new School, which has a large network
of computers and workstations for teaching and research (SUNs, HP
9000/300s, VAX, GEC-63, Orion-2 etc.) and many industrial connections.
Current research interests include language, vision, learning,
intelligent documentation tools, logics for AI, logic programming,
development of AI languages and tools (POPLOG development), computers
in Education, philosophical foundations of AI, AI and psychology,
computational linguistics.
The preferred start date is 1st October 1988, and applications should
be received by October 30th 1987, though later applications will be
considered.
For further information and application forms please apply to:
The Personnel Office
Sussex House
University of Sussex
Brighton
BN1 9RH, England
Phone: (+44) (0)273 - 606755
Aaron Sloman
Cognitive Sciences, Univ of Sussex, Brighton, BN1 9QN, England
Phone: University (44)-(0)273-678294
UUCP: ...mcvax!ukc!cvaxa!aarons
ARPANET : aarons%uk.ac.sussex.cvaxa@cs.ucl.ac.uk
JANET aarons@cvaxa.sussex.ac.uk
------------------------------
Date: 26 Sep 87 18:57:21 GMT
From: ihnp4!occrsh!erc3ba!erc3bb!cord!packard!edsel!granjon!io!mtunk!m
tune!whuts!homxb!homxc!del@ucbvax.Berkeley.EDU
Subject: Re: IEEE ASSP April 1987 on Neural Networks
I found the article very interesting and decided to code the hamming
classification neural network. I thought the comp.ai would be
interested. I think I found a bug in the article's description of the
routine (see code).
It should be portable to any C compiler.
_______________________cut here______________________
/* ham.c
* c version of the hamming net
* david leasure
* 9/25/87
*
* this routine is a hamming classification network
* described in IEEE ASSP April 1987 by Richard P. Lippmann pg. 9
* correcting for a presumed bug in the presented routine
* the bug is the value set for THETA by Lippmann. When THETA is
* N / 2 it so overwhelms the outputs from the lower net that only 0
* activation values are passed up from the threshold function.
* I have chosen to set epsilon to 1 / 2M and to not have an upper
* limit on the threshold function so no saturation occurs
*
* the program is somewhat inefficient because of the use of
* data storage for maxnet (t[k,l] in lippman's) and for output[t,M]
* but they could be useful in a simulator of this network which allowed
* things to be fiddled with.
* the code could be improved by not encoding the size and values
* of the node matrices directly, too, reading them instead from files
* and/or a user interface.
*
* if you improve the code, please send me the diff's
* david e. leasure
* ihnp4!homxc!del or del@homxc.att.com
*/
[Contact the author if you need the code. It's about
7000 characters. -- KIL]
--
David E. Leasure - AT&T Bell Laboratories - (201) 615-5307
------------------------------
End of AIList Digest
********************
