Copy Link
Add to Bookmark
Report
AIList Digest Volume 5 Issue 231
AIList Digest Monday, 12 Oct 1987 Volume 5 : Issue 231
Today's Topics:
Queries - Beginning Text for AI and LISP &
Beta Sites for Lucid Common Lisp for the Sun 4 &
Planning Knowledge and Representation &
Public Domain PROLOG via FTP & Common Loops &
Scheme on the SUN & Commercial Uses for Neural Nets &
Neural Network Texts,
Neural Networks - 500K Connections per Second,
AI Tools - CMU Common Lisp Distribution
----------------------------------------------------------------------
Date: 8 Oct 87 09:53:05 GMT
From: marshek@ngp.utexas.edu (MAt)
Subject: Re: A good beginning text for AI & LISP
Hi there,
***PLEASE E-MAIL RESPONSES TO ME***
I would like to know about a good introductory text for AI and
the LISP language. I have a background in C, fortran but I presume that
does not help much, does it?
Thanx in advance
MAt
------------------------------
Date: Thu, 8 Oct 87 09:47:47 PDT
From: edsel!sears@labrea.stanford.edu (Steve Sears)
Subject: Looking for beta-sites to test Lucid Common Lisp for the Sun
4.
Lucid Common Lisp for the Sun 4 will soon be here. We are
interested in locating sites that can provide good feedback by
exercising our Sun 4 beta-test version. If you are interested, or know
someone that may be, please reply. You can also reach me by phone at
Lucid, (415) 329-8400.
Thanks in advance...
Steve Sears
------------------------------
Date: 9 Oct 87 02:26:39 GMT
From: lukose@aragorn.cm.deakin.OZ (Dickson Lukose)
Reply-to: lukose@aragorn.OZ (Dickson Lukose)
Subject: Planning Knowledge & Representation Survey
G'Day Colleagues,
I'm a newcomer to the planning paradigm and AI in general. My
research interest is in planning heuristics. I'm currently doing
a survey on:-
(1) "planning knowledge",
(2) "knowledge representation for planning" and
(3) "process of planning"
used by various types (eg. nonhierarchical, hierarchical, script-based
and opportunistic) of planning systems.
(1) Is there anyone who have done the above survey or knows someone who
have done so?
(2) Is anyone aware of any survey publications related to the above
mentioned areas?
(3) Has anyone got any bibliography related to the above subject areas?
I'm most interested in communicating with researchers currently
involved in R&D of "planning systems".
Any suggestions or pointers to the above request will be much
appreciated.
If enough interest shown, I will transmit(e-mail) the results of survey.
------------------------------------------------------------------
Dickson Lukose | UUCP: ...!seismo!munnari!aragorn.oz!lukose
Div. Comp. & Maths | ....!decvax!mulga!aragorn.oz!lukose
Deakin University |
Victoria, 3217 | ARPA: munnari!aragorn.oz!lukose@SEISMO.ARPA
Australia | ACSNET: lukose@aragorn
------------------------------
Date: 8 Oct 87 17:47:44 GMT
From: aplcen!jhunix!ins_akml@mimsy.umd.edu (Katherine Martha Lai)
Subject: Public domain PROLOG available via ftp?
Can anyone tell me where there is a public domain PROLOG that I
could get via ftp? I would be running it on a Sun 3/160 under
UNIX. Thanks!!
I am posting this for a friend, so it probably would be best to
reply to him directly (Marty Hall) at "hall@hopkins-eecs-bravo.arpa".
------------------------------
Date: Fri, 9 Oct 87 14:59 EST
From: STREIFF%HARTFORD.BITNET@WISCVM.WISC.EDU
Subject: Common Loops Request
Hi,
Im looking for a copy of Common Loops. Does any one have a
copy or know where i can get one? Were running Lucid V1.2 on Sun3/75's.
Thank you.
+------------------------------+----------------------------------------+
| S. David Streiff | BitNet : STREIFF@HARTFORD.BITNET |
| CE-CIM-EE | SlowNet : Box 2590 200 Bloomfield Ave |
| Combustion Engineering | West Hartford, CT. 06117 |
| Windsor CT. | MaBell : (203) 726-9117 |
-------------------------------+----------------------------------------+
DisClaimer : My employer is not responsible for anything that i say.
They will even deny my existance if given a chance.
------------------------------
Date: 8 OCT 87 15:08-N
From: U00124%HASARA5.BITNET@WISCVM.WISC.EDU
Subject: Request: Scheme on the SUN?
Hello Scheme users!
Within short we hope to have a Sun 3/60 system considting of 4 units
backed by two 141 MB disks and a 60 MB tape streamer. The system comes
fully furnished: C, Fortran 77, Pascal and Modula-2 (and UNIFY).
We are interested in running Lisp, preferably Scheme because we use it
in different courses and we like the IBM-PC implementation!.
The question is: is there a implementation of Scheme for the SUN? and if
so: Where to get it, how much it costs, etc..?
Any other info concerning Lisp and Prolog on the Suns will be apreciated.
Dr. Oscar Estevez
Chairman Study Programme
Medical Informatics, Univ. of Amsterdam.
------------------------------
Date: 4 Oct 87 23:59:33 GMT
From: imagen!atari!portal!cup.portal.com!barry_night-person_stevens@uc
bvax.Berkeley.EDU
Subject: commercial uses for neural nets -- have, also need, info
I am working on a study of commercial uses for neural nets. So far, these
include processing commercial loan applications, insurance underwriting,
insurance claims processing, signature verification, face and/or voice
identification for security, financial optimization.
I am interested in swapping applications with those who know of others.
These applications are using both hardware (such as the Hecht-Nielsen ANZA
board) and software systems.
Please contact me by phone at 619-755-7231
or in writing:
Barry A Stevens
Applied AI Systems, Inc.
PO Box 2747
Del Mar, CA 92014
------------------------------
Date: 7 Oct 87 09:50:45 GMT
From: plx!titn!jordan@sun.com (Jordan Bortz)
Subject: Neural Networks - Pointers to good texts?
Hello;
I'm looking for some good texts and/or articles on neural networks,
in English, and preferably focusing on real life algorithms/implementations
rather than obscure mathematics.
If you know of any, please let me know by mail, and I'll summarize
to the net; as this topic seems to be generating more interest.
I know some articles were mentioned earlier, but what about others?
Jordan
--
=============================================================================
Jordan Bortz Higher Level Software 1085 Warfield Ave Piedmont, CA 94611
(415) 268-8948 UUCP: (decvax|ucbvax|ihnp4)!decwrl!sun!plx!titn!jordan
=============================================================================
------------------------------
Date: Thu 8 Oct 87 14:28:06-EDT
From: Dave.Touretzky@C.CS.CMU.EDU
Subject: 500K connections per second
The inner loop (and most expensive part) of neural net simulations computes for
all j the net input to unit j, which is the sum for all i of the output of unit
i times the weight Wji on the connection from i to j. This is just a multiply
and accumulate loop. In fact, if you choose the right data structures, it's a
matrix-vector multiplication. So when someone advertises that their
"neurocmputer" does 500K connections per second, they mean it does five hundred
fetch-multiply-accumulate operations per second. This is a useful performance
measure because it is independent of the number of units and connections in the
model being simulated.
There is, unfortunately, no such thing as a commercially available
neurocomputer. Presumably, a neurocomputer would either be a computer made out
of neurons, or a computer whose physical structure in some way resembled that
of the nervous system. No product available today meets either of those tests.
What people are selling today as "neurocomputers" are just regular old
computers with some neural net simulation software. For example, Hecht-Nielsen
NeuroComputer Corporation, the outfit that's been running those full-page
four-color ads in AI Magazine, sells their ANZA "neurocomputer" for $15,000.
The ANZA system is an off-the-shelf IBM PC/AT with an add-on board containing a
Motorola 68020 with floating point co-processor and 4Meg of memory. For
roughly the same price you could buy a Sun III (same 68020 processor) and run
Unix and X-windows instead of PC-DOS. In fact, Hecht-Nielsen will be
announcing a version of their simulation software for the Sun in the near
future. That doesn't make the Sun III a neurocomputer, but then again, neither
is the ANZA.
The TRW Mark III is also a coprocessor build out of conventional components,
but it attaches to a Vax rather than an IBM PC. The Science Applications
Corporation Sigma-1 is a high speed number cruncher based on a Harvard
architecture (the single processor has separate data and instruction paths); it
is not a neurocomputer. Science Applications recently acquired a Connection
Machine which they plan to use for really heavy duty simulations. (Connection
machines aren't neurocomputers either; they're much more general purpose than
that. See the article by Blelloch and Rosenberg in IJCAI-87 for a report on
using a CM2 to simulate learning in neural nets.)
The TI Oddyssey DSP (Digital Signal Processor) is another board that does fast
matrix-vector multiplies. Like the other products I mentioned, it is a
conventional architecture, basically a handful of TMS 98020(?) hardware
multiplier chips. I have a special fondness for Texas Instruments because even
though they do some interesting neural net research, they never use the
misleading term "neurocomputer" in their ads for the Oddyssey.
Will there ever be real neurocomputers? Perhaps some day:
Some people are building VLSI circuits whose structure is based on an abstract
description of neural circuitry. For example, a group at BELLCORE led by
Joshua Alspector and Robert Allen has designed a 54-unit "Boltzmann Machine"
chip. The 54 neurons are physically implemented as separate processors on the
chip, and their N*(N-1)/2 weighted connections are also implemented by separate
pieces of circuitry, giving a fully parallel implementation. This is terrific
work, but it will be quite a while before it has any commercial impact, because
it's hard to put a lot of neurons on one chip, and expensive to communicate
across multiple chips. It is possible to cram several hundred neurons on a
chip if you go for fixed weights (resistors) rather than variable ones, but
then the network can't learn.
Carver Mead and Mass Silviotti at Caltech have built a "silicon retina" low
level vision chip using analog (!) VLSI circuitry. The chip's architecture was
inspired by the way real retinas do computation.
There is also work on optical implementations of neural networks, using lasers,
two-dimensional or volume holograms, and various mirrors and photosensors. Two
of the big names in this area are Dmitri Psaltis (Caltech) and Nabil Farhat
(Penn). It will probably take longer for this technology to reach the
marketplace than for VLSI-based technologies, as it is in a much earlier stage
of development.
A group at Bell Labs has been growing real neurons on a special substrate with
embedded electrodes, so they can have an electronic interface to a living
neural circuit. This is a neat way to study how neural circuitry works, but
they only deal with a handful of neurons at a time. I doubt whether it will
ever be practical to design special-purpose computers from living neurons.
A good place to learn more about neuromorphic computer architectures (a more
decorous term than "neurocomputer", in my opinion) is the proceedings of neural
net conferences. There's the proceedings of the 1986 Snowbird Meeting on
Neural Networks for Computing, published by the American Institute of Physics
in New York. There's also the IEEE First International Conference on Neural
Networks, which was held in San Diego this past June. And there's the IEEE
Conference on Neural Information Processing Systems - Natural and Synthetic,
which will be held in Denver, at the Sheraton Denver Tech Center, on November
8-12. This conference was originally to take place in Boulder, but
registration has been so heavy it had to move to larger quarters at the last
minute. The conference chairman is Yaser Abu-Mostafa at Caltech.
Sorry, I don't have information on how to order proceedings from the IEEE
conferences. Contact the IEEE.
-- Dave Touretzky
------------------------------
Date: Mon, 05 Oct 87 11:41:30 PDT
From: Vicki L. Gordon <vgordon@venera.isi.edu>
Subject: CMU Common Lisp Distribution
With DARPA funding, the University of Southern California's Information
Sciences Institute (USC/ISI) is serving as a distribution center for
public-domain Common Lisp software packages. The first of these packages
includes the source files for CMU Common Lisp (formerly known as "Spice
Lisp"). The package also includes an unsupported public-domain version
of the OPS5 language that is written in Common Lisp.
If there is sufficient interest, and if funding allows, we will later expand
the distribution program to include additional software packages from other
sources.
CMU Common Lisp is a full implementation of Common Lisp, developed as part of
the Spice project at Carnegie-Mellon University (CMU). This system runs only
on the IBM RT PC, and only under CMU's Mach operating system (superficially
similar to 4.3bsd Unix, but with a different internal organization). Since
the IBM RT version of Mach is not currently supported outside of CMU, it
follows that CMU Common Lisp is *NOT* a system that anyone can obtain and
run as is. We are making the sources for this system available because a
number of groups have found it to be a useful starting point for their own
Common Lisp implementations. Typically, it takes a man-year of effort to port
this system to a new machine and operating system (more if the target
environment is unusual). Individuals and small research groups who want a
Common Lisp for their own use are advised to use one of the commercially
available products.
This source code is in the public domain and, once you have them, there
is no restriction on how you may use them. They are made available as a public
service, with no warranty of any sort by the authors, CMU, or USC/ISI, and
with no promise of future support. CMU Common Lisp has been heavily used at
CMU, but it has not been extensively tested in any systematic way. Questions
about the distribution procedure may be directed via electronic mail to
ACTION@ISI.EDU, or you may call (213) 822-5511. Bug reports and questions
about the code itself should also be directed to SCOTT.FAHLMAN@CS.CMU.EDU.
This distribution package contains approximately 7 megabytes of ASCII source
files. It can be obtained over the Arpanet/Milnet by establishing an FTP
connection to VENERA.ISI.EDU and logging in as "anonymous" (any password can be
used). For optimal response time, please conduct your file transfer during
our non-primetime hours (1800 to 0800 PDT). The source files are kept in the
following directories:
/common-lisp/implementation/cmu/code
(Runtime system and interpreter for CMU Common Lisp on IBM RT PC under Mach)
/common-lisp/implementation/cmu/clc
(Compiler from Common Lisp to native code for RT/Mach. Mostly written in
Common Lisp itself.)
/common-lisp/implementation/cmu/hemlock
(Sources for the Hemlock text editor. This editor is similar at user level to
Emacs. Written in Common Lisp, but contains some Mach-specific system calls
and display code. Uses either X windows or standard termcap terminals.)
/common-lisp/implementation/cmu/OPS5
(Portable Common Lisp version of the OPS5 production-system language.
A quick and dirty port of the public domain version that was developed
originally in Franz Lisp. No support whatsoever is provided.)
/common-lisp/implementation/cmu/miscops
(Low level support routines for the IBM PC RT: garbage collection, generic
arithmetic, etc. Totally machine-specific. Provided as an example of what
needs to be done in order to port this lisp to another machine.)
/common-lisp/implementation/cmu/icode
(Lisp functions implementing the interface between Mach and CMU Common Lisp.
Very system-specific. Provided only as an example.)
/common-lisp/implementation/cmu/lib
(Cursor definition file and spelling dictionary used by the Lisp runtime
system).
If you would like to order a tape, we will first send you a release form
which you are required to sign prior to receiving the tape. When you return
the signed release, include a check made payable to USC/Information Sciences
Institute for $100.00 to cover the production and mailing costs. Please
remember to include a complete return address.
The default tape format will be tar 1600 bpi, unless otherwise specified.
Currently ISI is only prepared to distribute tapes containing the CMU
Common Lisp code to individuals or companies within the United States.
We are currently negotiating with the Department of Commerce and CMU
to obtain authorization to distribute the code to countries outside of
the United States; however, we do not expect approval in the immediate
future.
The following hardcopy documentation produced by CMU is also available
to all recipients at a cost of $20.00 per package (payable to USC/
Information Sciences Institute). The package includes:
- "Internal Design of Common Lisp on the IBM RT PC"
- "CMU Common Lisp User's Manual, Mach/IBM RT PC Edition"
- "Hemlock User's Manual"
- "Hemlock Command Implementor's Manual"
Please send your request for a tape and/or documentation to ACTION@ISI.EDU,
or mail it to the following address:
USC/Information Sciences Institute
Attn: ACTION
4676 Admiralty Way
Marina del Rey, CA 90292
(213) 822-1511 ext 289
------------------------------
End of AIList Digest
********************
