Copy Link
Add to Bookmark
Report
AIList Digest Volume 6 Issue 040
AIList Digest Saturday, 27 Feb 1988 Volume 6 : Issue 40
Today's Topics:
Adminstrivia - Head Count Results & Speaker's Net Addresses,
Correction - Lotfi Zadeh's Nationality,
References - 1988 Canadian Artficial Intelligence Conference &
Schank's Papers & Constraint Satisfaction Programming,
AI Tools - Neural-Net Simulators & Fuzzy Logic and Probability Theory,
References - Introduction to Parallel Processing
----------------------------------------------------------------------
Date: Fri 26 Feb 88 11:46:02-PST
From: Ken Laws <LAWS@SRI.COM>
Subject: Head Count Results
I received 30 responses to my request for messages from
readers with birthdays in early February. (There were 19
from the U.S., 3 each from Australia and the UK, 2 from
Canada, and 3 from Bitnet sites unknown to me.) Multiplying
by the appropriate time-span factor (365/15) shows that there
were about 730 readers alert, able, and willing to reply
to the request.
The number of AIList readers is obviously much higher than
that. Bitnet alone distributes to 400 addresses (only a
few of which are known to be further redistributions).
Applying this factor of 3/400 to the full response of 30
implies that there are about 4000 readers physically able
to reply. (Many readers, e.g. in Great Britain, do not
have outgoing mail priviledges due to the expense involved.
Many others in Europe, Japan, South Korea, and elsewhere
may be unable to construct the necessary return path.)
Another handle on the readership is the number of AAAI
members with net addresses. I don't have access to the full
membership list, but about 3500 of the members have net
addresses short enough to list in the printed directory.
If we assume that 3000 such members in the U.S. are all
Arpanet/CSNet AIList readers, the full 30 replies would
represent a pool of 4737 AIList readers worldwide. This
is certainly imprecise, but perhaps U.S. net members not
reading AIList are balanced (or more than balanced) by
students reading AIList bboards and by readers with long
net addresses (when viewed from the Arpanet).
My conclusion is that there are probably around 4000
readers (perhaps 3000 to 6000), with fewer than 1000
able to respond to any given query or message. A more
precise (or accurate) estimate of the readership would
require surface mail as the reply medium.
-- Ken
------------------------------
Date: Tue, 23 Feb 88 14:10 N
From: MFMISTAL%HMARL5.BITNET@CUNYVM.CUNY.EDU
Subject: SPEAKER'S NET ADDRESSES IN SEMINAR ANNOUNCEMENTS
The seminar announcements on AIlist digest would be even more helpful
especially for those on the other side of the ocean, when the
net addresses of the speakers would be included. This would facilitate
us with a direct way to ask for more information such as reports.
I normally send an EMAIL to the organizers, but a direct message to the
speakers would be easier.
Jan L. Talmon
Dept of Medical Informatics and Statistics
University of Limburg
Maastricht, The Netherlands
EMAIL: MFMISTAL@HMARL5.Bitnet
------------------------------
Date: Tue, 23 Feb 88 17:29:47 EST
From: ST401843%BROWNVM.BITNET@MITVMA.MIT.EDU
Subject: Mea Culpa
As many people pointed out to me, Lotfi Zadeh is an Iranian (Persian as
we used to say in the olden days) and not an Indian. Mea Culpa!!!
------------------------------
Date: 21 Feb 88 16:36:13 GMT
From: mnetor!utzoo!utgpu!jarvis.csri.toronto.edu!ai.toronto.edu!gh@uun
et.uu.net (Graeme Hirst)
Subject: Re: 1988 Canadian Artficial Intelligence Conference
In article <8802111540.AA20271@jade.berkeley.edu> ABCANO01@ULKYVX.BITNET writes:
>"Submitted to the 1988 Canadian Artificial Intelligence Conference."
>Can anyone tell me anything about this conference, esp. dates, costs, etc?
The Canadian AI conference is in Edmonton, Alberta, the week of June 6-10,
in conjunction with the 1988 Canadian vision and graphics conferences.
For general information, write to
Wayne Davis, Conference Chairman
Dept Computing Science
University of Alberta
Edmonton, Alberta
CANADA T6G 2H1
Phone: 403-432-3976
--
\\\\ Graeme Hirst University of Toronto Computer Science Department
//// utcsri!utai!gh / gh@ai.toronto.edu / 416-978-8747
------------------------------
Date: 22 Feb 88 06:04:38 GMT
From: woodl@byuvax.bitnet
Subject: Re: We look for Schank's papers
In response to the search for references on Roger Schank's work,
Can I assume you are aware of Schank's article in the latest (Winter
1987)
issue of AI Magazine entitled "What is AI, anyway?" It lists many of his
published references plus Yale Technical Reports. The article following it
is a report on the AI efforts at Yale, which also includes references by
Schank. If that doesn't help, I can make up a list of important ones and
send to you.
Larry Wood, Brigham Young University, WoodL@BYUVAX.bitnet
------------------------------
Date: 25 Feb 88 18:17:16 GMT
From: zodiac!sun-oil!rlee@spam.istc.sri.com (Richard Lee)
Subject: Re: constraint satisfaction programming
In article <5070@pyr.gatech.EDU> parvis@pyr.gatech.EDU (FULLNAME) writes:
<I saw an excellent talk a couple of years ago by a man named Wm. Leler
<in which Wm. (pronounced Wim) discussed a constraint language called
<Bertrand. This language was developed by Wm. in connection with research
<leading to his Ph.D. His Ph.D. thesis has since been published as a
<distinguished thesis by one of the computer science publishers. It's called
<Constraint Languages. It's fairly recent. Don't know the publisher right
<off hand.
<
<Phil Miller
_Constraint Programming Languages: Their Specification and Generation_,
by Wm Leler, Addison-Wesley, 1988, ISBN 0-201-06243-7.
Wm is short for William.
------------------------------
Date: Mon, 22 Feb 88 09:32:28 PST
From: heirich@cs (Alan Heirich)
Subject: Re: AIList V6 #38 - Applications, Neuromorphic Tools, Nanotechnology
AILIST #38 contained a summary of sources of neural network software.
They missed two important packages used in academic research:
* The Rochester Connectionist Simulator, available from the Computer Science
Department of the University of Rochester. A modifiable package, written
in C. Allows interactive design and use of neural networks using a variety
of learning algorithms. New algorithms can be easily coded in C.
* SunNet, available from the Institute for Cognitive Science, University of
California San Diego. A "closed" package, it includes a simple programming
language that allows procedures to be written to implement learning
algorithms. Oriented toward back propagation learning, but can be extended
to other types. Superb graphics and very easy to use.
I believe that both of these packages are available for a nominal cost which
covers media and handling. The Rochester simulator includes source code; I
don't know about SunNet, but it may well include the source.
- Alan Heirich (heirich@sdcsvax.ucsd.edu)
------------------------------
Date: 22 Feb 88 05:25:05 GMT
From: quintus!ok@sun.com (Richard A. O'Keefe)
Subject: Re: I'm still not convinced ... Fuzzy Logic and Probability
Theory
I don't like "fuzzy logic". The basic reason for that is very simple:
it puts the fuzziness in the wrong place. The standard example is
"John is very tall"
where "very" is interpreted as a degree-of-belief in the proposition
"John is tall".
This fails to make it clear whether
- there is some doubt about the height
or - someone is tall, but there is some doubt about whether it's John.
This ambiguity does not exist in the original statement. Putting the
fuzziness on the truth-values instead of the functions seems wrong.
In probability, there is a clear distinction between distributions
(relating to functions) and probabilities (relating to propositions).
------------------------------
Date: 23 Feb 88 07:27:37 GMT
From: calgary!gaines@sri-unix.ARPA (Brian Gaines)
Subject: Re: FUZZY LOGIC VS. PROBABILITY THEORY
In article <8802180658.AA11175@ucbvax.Berkeley.EDU>, golden@FRODO.STANFORD.EDU
(Richard Golden) writes:
> The basic theoretical result is that selecting a "most probable" conclusion
> for a given set of data is the ONLY RATIONAL selection one can make in
> an environment characterized by uncertainty. (Rational selection in this
> case meaning consistency with the classic deductive/symbolic logic - boolean
Boolean logics are not appropriate for knowledge representation if the
underlying domain is truly fuzzy, that is, has borderline case where either,
x and not x are both true, or, x and not x are both false. A classic example
is shades of color that grade into one another.
> algebra.) Thus, one could argue that if one constrains the class of
> possible inductive logics to be consistent with the laws of deductive logic
> then Probability Theory is the MOST GENERAL type of inductive logic.
>
> (iii) F(C and B,A) may be computed from F(C,B and A) and F(B,A)
> Note this assumption's similarity to Bayes Rule but the
> multiplicative property is not assumed.
This is an assumption of truth functionality that is excessively strong.
In general we cannot infer truth values for conjunctions in this way.
>
> To my knowledge, the axioms of Fuzzy Logic can not be derived from
> consistency conditions generated from the deductive logic so I conclude
> that Fuzzy Logic is not appropriate for inferencing. Any comments?!!!
>
There is a strong relation between classical, probability and fuzzy logics.
If one starts with a lattice of propositions and an additive measure over it
such that p(a and b) + p (a or b) = p(a) + p(b), then one gets a general
logic of uncertainty that:
a) Becomes probability logic if you assume excluded middle, ie no borderline
cases;
b) Becomes fuzzy logic if you assume strong truth functionality, ie p(a and b)
can be inferred from p(a) and p(b);
c) Becomes standard propositional logic if you assume binary truth values.
Most of the useful results in fuzzy and probability logics can be derived
in the general logic and do not need the restrictive assumptions.
Theorem provers for any of the multivalued logics are essentially
constraint chasers that bound the truth values of propositions, more
like linear programming than classical resolution.
There is a wealth of literature on fuzzy and probability logics -
The journals Fuzzy Sets and Systems, Approximate Reasoning, and Man-Machine
Studies, all carry articles on these issues and applications to
knowledge-based systems. Noth-Holland have published several books
edited by Gupta, Kandel, Yager, and others, on these topics.
Brian Gaines, gaines@calgary.cdn, (403) 220-5901
------------------------------
Date: 25 Feb 88 00:56:40 GMT
From: pyramid!leadsv!esl!ian@hplabs.hp.com (Ian Kaplan)
Subject: Re: Parallel Processing
jb@otter.hple.hp.com (Jason Brown) writes:
>
>To the world,
>
>I have just read a book entitled "Supercomputers of today and tomorrow the
>parallel processing revolution", by richard a jenkins. (TAB press). This book
>goes through the basics of parallel processing and what is happening in the PP
>world at the moment. The book was wriiten in 1985 and is some what put of data
>, I should imagine.
> [ text deleted ]
>
>It would be nice if I could contact the originators [of the parallel
>machines]>, so if anyone reads this who is at any of the above places
>and knows, or can find out where the person is maybe you could get
>them to contact me.
>
Jason:
I could not get your address to word via e-mail, so here is my note.
If you want a more detailed survey read the book by Hwang and
Briggs. Its title is something like "Parallel Processor
Architecture". I think that it was published by McGraw-Hill.
You are rather naive in your belief that someone is going to send
you a note explaining all the trade offs in the parallel architecture
they chose. Such an explaination would be more appropriate as a
seminar or course, rather than a short note. Also, why should
researchers spend their time when you will not even spend the time to
go to the library and read more extensively on the topic. Tab press
is infamous for printing trash. You also asked for references. Even
a reference list would be extensive. However, a technical library
should have some bibliographies.
The big conference on parallel processing is the International
Conference on Parallel Processing. Look at the conference
proceedings. The Proceedings of the International Symposium on
Computer Architecture also has articles on parallel processings on
occasion.
Finally, take a class on computer architecture. You cannot understand
the tradeoffs in parallel architectures unless you understand standard
serial and vector architectures.
Ian L. Kaplan
ESL, Advanced Technology Systems
M/S 302
495 Java Dr.
P.O. Box 3510
Sunnyvale, CA 94088-3510
esl!ian@ames
decvax!decwrl!\
sdcsvax!seismo!- ames!esl!ian
ucbcad!ucbvax!/ /
ihnp4!lll-lcc!
------------------------------
End of AIList Digest
********************
