Copy Link
Add to Bookmark
Report
AIList Digest Volume 8 Issue 058
AIList Digest Friday, 19 Aug 1988 Volume 8 : Issue 58
Query Responses:
A public-domain computer chess program
Camera Stabilization
Looking for a Cognitive Science Society
Garden Design and Plant diagnosis
Sigmoid transfer function
Feigenbaum's citation
----------------------------------------------------------------------
Date: Sun, 14 Aug 88 11:45:47 -0200
From: Antti Ylikoski <ayl%hutds.hut.fi%FINGATE.BITNET@MITVMA.MIT.EDU>
Subject: a public-domain computer chess program
In AIList Digest V8 #31, Rohit Gupta
<uxe.cso.uiuc.edu!gupta@uxc.cso.uiuc.edu> writes:
>I will be starting my Master's this fall and am fascinated by Artificial
>Intelligence - especially in computer chess.
>Does anyone know of any good info (books, papers, authors, professors,
>articles, research projects) on this subject?
The magazine Creative Computing published a large (several thousand
lines long) chess program written in Pascal and running in a large Cyber
computer I think in the end of the 70's or in the beginning of the 80's.
I recall the article and the program were written by one of the famous
computer chess people, possibly by Hans Berliner.
--- Andy
Disclaimer: the writer of this entry likes to give the impression of
being more intelligent than he is, is known to have written an AIList
entry after having several beers and even is a member of the Mensa.
------------------------------
Date: 15 Aug 88 14:27:00 GMT
From: aplcen!jhunix!apl_aimh@mimsy.umd.edu (Marty Hall)
Subject: Camera Stabilization
In article <49@cybaswan.UUCP> eederavi@cybaswan.UUCP (f.deravi) writes:
>I am looking for information on camera stabilization and sensors for
>this purpose suitable for moving vehicles.
The Robotics group at AAI Corp here in Baltimore build gyro-stabilized
gimballed mounts for various cameras and sensors. My understanding
is that there are aluminium and carbon composite versions, and are
suitable for either ground or air vehicle use.
You can contact Steve Moody for more info:
Mr. Steve Moody
Robotic Systems Operations
AAI Corporation
PO Box 126
Hunt Valley, MD 21030 USA
(301) 628-3189
------------------------------
Date: 15 Aug 88 15:42:58 GMT
From: trwrb!ries@bloom-beacon.mit.edu (Marc Ries)
Subject: Camera Stabilization
[...]
>
> Panasonic showed a gyroscopically stablized consumer-grade camcorder
>at the Consumer Electronics Show this summer. It should be available at
>Japanese retailers by the end of the year. This may be a promising approach.
>
I believe that either VIDEO or V. Review has a usage report on
the new Panasonic/Mitsubishi stabilized camcorder this month.
Negatives: Price ($2400), weight (9+ pounds) and format (VHS only)
Pluses: It works.
--
Marc A. ries@trwrb.TRW.COM
sdcrdcf!---\
ihnp4!------\----- trwrb! --- ries
------------------------------
Date: Mon, 15 Aug 88 09:33:19 PDT
From: norman%ics@ucsd.edu (Donald A Norman-UCSD Cog Sci Dept)
Reply-to: danorman@ucsd.edu
Subject: Looking for a Cognitive Science Society
Of course there is a cognitive science society-- ten years old this
year.. Just go to the library and look for Cognitive Science, the
journal -- published 4 times a year. This will also have the name and
address of the society. And, yes, it has an annual convention, in
Montreal this year right this very moment, almost: Aug 17, 18, and 19.
The current secretary/treasurer is Kurt VanLehn
Kurt VanLehn
Department of Psychology
Carnegie-Mellon University
Pittsburgh, PA 15213
vanlehn@a.psy.cmu.edu
But his term is now over and a new person will be selected at the
Montreal meeting.
don norman
Donald A. Norman
Department of Cognitive Science C-015
University of California, San Diego
La Jolla, California 92093
INTERNET: danorman@ucsd.edu INTERNET: norman@ics.ucsd.edu
------------------------------
Date: Mon 15 Aug 88 11:51:28-PDT
From: Leslie DeGroff <DEGROFF@INTELLICORP.COM>
Subject: Garden Design and Plant diagnosis
Agricultural and plant diagnostic sytems reply
For those working on or exploring garden design systems (a while ago)
or plant diagnostic systems, I would suggest two things,
1. Just to start, expand your search for info to include DATABASES. A number
of things have been done with database tools that can in effect provide
a ready made knowledge base. One potential contact (start point on
a completely different intellectual network) would be
Dr. Bashem,
College of Agriculture, Colorado State University,
Ft Collins CO.
He and others at CSU have built an ornamentals (flowers, trees but not food
plants) database of several thousand plants and about 40 fields including
botanic and common names, growth habits and cultivation requirements. They
had fields for common disease and insect problems but they are sparsely
filled. This database was originally built in RIM (mainframe database tool
from Boeing) and was being ported to RBASE (tm) in 1986. RBASE is available
on a variety of micro platforms and has a (sold separate) program interface
library.
Also going back to 1986, I saw a very limited (almost toy) commercial
equivalent with 700+ plants and <20 feilds with access software being
sold by Ortho (I think) It was cheap, on order of 40$. No programmatic
interface.
2. Having reoriented toward databases, you may want to pursue the
search by contacting (in the US) the Cooperative Extension Service,
Agricultural Research Service and state Agricultural Colleges. If you
have a particular crop in mind, it is likly that someplace in the US there
is a researcher who has spent his life on it. The ARS and Cooperative
Extension are government agencies and are chartered to do and
collect research information and to provide this information to
farmers and others needing it. (a major goal is disemination of
working solutions to producers and consumers) I would suggest
treating this as an information network, if the individuals you
talk to don't have the information you want, ask if they can
point you to someone else. For the actual knowledge to go into
a plant diagnostics system, you almost certainly will be back to
research and researchers involved in this system, but if you
search for databases you may find a massive amount of work already done.
It is also true some of those databases are being used like
limited Expert Systems, with the ornamentals database at CSU, one or two
hours of training could teach Horticultural Design students how to set up a
query to create a list of plants for a desired situation.
I would suspect that there are several ES projects underway within the
ARS, Extension and agricultural college systems, without having
any specific pointers, I would suggest contacting departments at
Purdue and Texas A&M they have been active in related areas, Texas was
working on an impressive demonstration farm sensor and automation project
in 1984.
Relevant departments would include:
Horticulture (landscape, flowers and vegetables)
Agricultural Engineering (often building sensors and monitoring systems
for other research groups in college)
Agronomy (field crops like corn, wheat, cotton)
Plant Pathology
Entomology
Leslie DeGroff (DeGroff@Intellicorp.Arpa)
(Agriculture is the root of civilization)
------------------------------
Date: 16 Aug 88 06:05:00 GMT
From: a.cs.uiuc.edu!uicslsv!bharat@uxc.cso.uiuc.edu
Subject: Sigmoid transfer function
erf(x)
Eqn. A
/ X
erf(X) = | e ( - X*X/(sqrt 2pi)) dX
| __________________________
/(- inf) (sqrt (2 pi))
In some references the erf is defined as integral from {0} to {X}
rather than {-inf} to {X}.
However if you do not wish to store a table of values of erf, then
numerical methods can be employed to compute the erf to a desired
precision. The following formula may be used.
Eqn B.
/ Y
| e ( - X*X/(sqrt 2pi)) dX
| __________________________ = (erf(Y) - 1/2)/(sqrt (2 pi)) =
/ 0
+inf (i-1)
_____ ______
\ | |
\ j >=0| |(2j+1)
/ ___________________ (y ** (2i+1))
/
_____ (2i + i)!
i >=0
The degree of precision that is required determines the number of
terms from the series that are needed to ensure convergence. If less
than 2&1/2% error is sufficiently precise, it can be assumed that the
erf(x) approx 1.0 for deviations from the mean (mu) greater than twice
the standard deviation (sigma). In that case the first 4 terms from
Eqn B are sufficient to lead to an accurate result. If greater than
!/2 % accuracy is required, then atleast the first 7 terms of the
series must be considered, and you can assume that erf(x) approx 1.0
for x- mu >=3 sigma, and erf(x) approx 0.0 for x - mu<= 3 sigma .These
approximations can then be incorporated into Eqn 4.2 to facilitate
speedy calculation.
(For the above formula Y = (x-mu)/sigma, or for any distribution with
mu = 0, and sigma = 1.
-Bharat
R.Bharat Rao
bharat%uicsl@uxc.cso.uiuc.edu
bharat@uicsl.csl.uiuc.edu
------------------------------
Date: 16 Aug 88 23:48:19 GMT
From: beowulf!pluto@sdcsvax.ucsd.edu (Mark E. P. Plutowski)
Subject: Feigenbaum's citation
It is interesting that since Japan has been quiet about
their progress (upon the Fifth Generation Project) it is
assumed that they have therefore progressed very little.
Now, I might assume this about American-based companies,
especially publicly owned ones. But is this true in Japan?
Does anyone know the facts here?
[Aside: when i read Feigenbaum's book when it came out,
just a few years earlier Japanese products were the butt of
jokes. Now, American products are.
(as reported in one of the business trade journals about
the increasing number of Americans working for Japanese
managers. according to the article, Japanese managers consider
Americans "lazy and untrustworthy.")]
Don't flame me, I bought an American car. But, isn't their
track record good enough of late to take their even
most ambitious plans seriously?
----------------------------------------------------------------------
Mark Plutowski
Department of Computer Science, C-024
University of California, San Diego
La Jolla, California 92093
INTERNET: pluto%cs@ucsd.edu pluto@beowulf.ucsd.edu
BITNET: pluto@ucsd.bitnet
UNIX: {...}!sdcsvax!pluto
------------------------------
Date: 17 Aug 88 02:58:34 GMT
From: cck@deneb.ucdavis.edu (Earl H. Kinmonth)
Subject: Re: Feigenbaum's citation
In article <5226@sdcsvax.UCSD.EDU>
pluto@beowulf.UUCP (Mark E. P. Plutowski) writes:
>It is interesting that since Japan has been quiet about
>their progress (upon the Fifth Generation Project) it is
>assumed that they have therefore progressed very little.
They haven't been silent. They publish annual and other reports.
I've gone through several with Japanese engineering friends
looking for content. There wasn't much. Playing it close to the
vest is NOT Japanese style for show-piece projects like this. If
they had something, they'd be crowing.
>Now, I might assume this about American-based companies,
>especially publicly owned ones. But is this true in Japan?
See above. Note that the fifth generation Project is not a
company in the conventional sense.
>Does anyone know the facts here?
>[Aside: when i read Feigenbaum's book when it came out,
>just a few years earlier Japanese products were the butt of
>jokes. Now, American products are.
You must have been living in a very rural area. Feigenbaum's
book was published in 1983. The Japanese reputation for quality
was well-established by the mid-1960s in general, and earlier for
products such as watches and cameras. I would say that the
Japanese reputation for quality was generally established two
decades before Feigenbaum published, except possibly for real
redneck areas of this country....
>(as reported in one of the business trade journals about
>the increasing number of Americans working for Japanese
>managers. according to the article, Japanese managers consider
>Americans "lazy and untrustworthy.")]
Public opinion polls in Japan show the Japanese think rather
highly of themselves. A more accurate generalization would be
that a good percentage of the Japanese consider all non-Japanese
lazy and untrustworthy....
>Don't flame me, I bought an American car. But, isn't their
Sympathy, yes. Flames, no.
>track record good enough of late to take their even
>most ambitious plans seriously?
No. Japan has it share of hucksters, con-artists, research
projects to which Proxmire would give his Golden Fleece Award,
and failures. Just because certain aspects of the economy are
doing exceptionally well should not lead to a "halo effect" that
blinds observers and causes them to abandon all serious
analysis. To do so would be to apply to Japan the same
uncritical approach Americans have tended to take with respect to
this country, especially in the 1950s and early 1960s.
"Ambitious plans" in Japan should be examined just as critically
as "ambitious plans" in the US. More bucks, more bull is a rule
that has equal applicability in both cultures. The history of
American writing on Japan (something I've taught as a course) has
shown one constant: wild exaggeration, whether the stereotype was
negative or positive.
------------------------------
End of AIList Digest
********************
