Parallel Processing Super Computer
I propose the following project:
Write a simple java applet. Get a number (5, 10, 100, 1000) peop=
le
on the net together. We all
log on to the net at the same time and run the applet. Each app=
let
has two states; on (1) and off (0).
Initially, everyone is off (0). The =93start host=94 initiates the ex=
periment
by manually changing his/her
system's status from 0 to 1.
The applet does the following simple tasks:
It reports a change in the state of its host machine to two other host=
s
by sending a change bit. The
receiving hosts for each participant are assigned by ordering the par=
ticipants
and assigning the next
two on the list to each host in the list. (see graphic below)
At the bottom, you loop up top and close
the system. (So the last host on the list would sen=
d
its change bit to the first two) [Eventually, it
would nice to automate this process with a script on a server to acco=
mmodate
ever larger
populations; but that can come later] The change bit could be
anything, say the number 1. The only
time the computer sends this information is when its status is change=
d
by receiving a change bit.
(That is another thing the JAVA applet does) It will receive
change bits from two other computers
in the network. When it receives a change bit, it changes the
current status from 0 to 1 or 1 to 0.
(Obviously, if every host SENDS change bits to two hosts, the
hosts must also accept change bits.)
The acceptance of a change bit merely switches the status of the accep=
ting
host regardless of the
status of the sender. That change in the accepting host trigger=
s
a SEND and it sends out a change
bit to the two computers it is set to send to.
The last thing the applet on each host does is records the status of
the host computer (0 or 1) at
some time interval, say every 30 seconds. We will have to exper=
iment
with different time intervals (it
may ultimately be much shorter than 30 seconds). For now, consi=
der
the time interval as the "scale".
The applet can be set to end after a specified time period. Say,
4 hours. That would mean for each
host there would be 480 status entries (one every thirty seconds).
Applet Inputs: 1. Two IP addresses of machines it
will send to. These could be sent by the
&nb=
sp; &nbs=
p;
receivers before starting.
&nb=
sp; &nbs=
p;
[That is, each participant sends their IP address to 2 people and receive=
s
IP
&nb=
sp; &nbs=
p;
addresses from 2 people.]
&nb=
sp;
2. Time interval for recording status (e.g. 30 seconds)
&nb=
sp;
3. For the start host, a way to manually change the status.
Applet Outputs: Status vector
Applet function: Hold current status in a variable
&nb=
sp; &nbs=
p;
Change current status when change bit is received
&nb=
sp; &nbs=
p;
Send change bit when status changes (i.e. change bit is received)
&nb=
sp; &nbs=
p;
Record status at predetermined time interval
The final step would be to gather each of the hosts status vectors int=
o
columns and graph the results.
The results would be graphed by assigning a black dot to a host that
was off and a white dot to a
host that was on. The column vectors would be lined
up from left to right based on the initial list
order.
The resulting graphic would be 100 pixels wide (if we had 100 particip=
ants)
and 480 pixels tall (for
each status check). I have some ideas about some fancier graphs with
more colors which I will share
if this thing gets off the ground.
If we could get this to work, it would be just a simple beginning.&nbs=
p;
Interaction between hosts could be
whatever we wanted (or could program).
Two exciting things about this model:
1. It is a Complex System.
2. It will run in parallel.
If participating in this would interest you, send me email at pberr=
y1@flash.net
by responding to this
post. I will gather up all the responses and let each responden=
t
know the status of the project. I am
thinking that initially we would want to stay below 100. Conver=
sely,
we could go forward with
anywhere from 5 to 10 participants. Once I get a handle on how
many folks want to participate, I
will send out my take on functional methodology such as assigning rec=
eiving
hosts, scheduling run
times, processing and distributing the results, etc.
Finally, the Java applet I described above is relatively straight forw=
ard.
Unfortunately, I was one of
those folks who got real excited about Java two years ago but any pro=
gram
skills I had developed
have deteriorated with lack of use. I would be willing to step
up and give it a shot but it would take
me a while. I figure there has got to be someone out there who
could knock it out pretty quick. Let
me know if you are such a person and would be willing to do it.
One last thing: The experiment outlined above is wide open. =
;
If you are interested in what I have
proposed, feel free to offer suggestions or changes. My only
goal is to actually try it. Keep in mind
that the first attempt should be relatively simple. Once
we do it once, then we can get fancy.
If you have any questions, please feel free to email me.
Pete Berry
--------------E8427C55921E7D482A05DA75
Content-Type: image/jpeg
Content-ID: CALL FOR PAPERS
------=_NextPart_000_002C_01BD1B88.748A6110--
From owner-population-bio@net.bio.net Sat Jan 10 22:00:00 1998
Path: biosci!internet!biosci!not-for-mail
From: biohelp (BIOSCI Administrator)
Newsgroups: bionet.population-bio
Subject: BIOSCI/bionet miniFAQ & Fundraiser
Date: 11 Jan 1998 02:00:11 -0800
Organization: BIOSCI International Newsgroups for Molecular Biology
Lines: 233
Sender: daemon@net.bio.net
Distribution: world
Message-ID: <199801111000.CAA28744@net.bio.net>
NNTP-Posting-Host: net.bio.net
(LAST REVISION: 30-JUL-95)
This BIOSCI "miniFAQ" is designed to answer the questions that come up
the *most frequently*. The main BIOSCI FAQ (Frequently Asked
Questions) is accessible on the World Wide Web at URL
http://www.bio.net/.
If you can not find an answer to your question in this or other
documentation, the BIOSCI technical support staff answers e-mail
queries sent to
biosci-help@net.bio.net
We can only answer questions about the use of the newsgroups and
mailing lists. We unfortunately do not have the staff to do Internet
information searches or answer scientific questions. Please post
those to the appropriate BIOSCI/bionet newsgroups.
Contents:
--------
0) BIOSCI NEEDS YOUR SUPPORT!!
1) Using the WWW to access the BIOSCI/bionet newsgroups.
2) What to do about "spams," i.e., junk mail, ads, etc.
3) Examples of subscribing and unsubscribing to the mailing lists.
4) The BIOSCI user address and research interest directory.
0) BIOSCI NEEDS YOUR SUPPORT!!
------------------------------
BIOSCI's government funding has been expended, and we are now
operating solely from advertising revenue that we have raised from our
Web site at http://www.bio.net/. We need just a few minutes of your
time to help us serve you.
You can do two important things which will take very little time for
you individually and will immensely help us continue to help you.
First, please use our WWW system at http://www.bio.net/ to access the
archives. You can post or reply to messages via your Web browser as
described in item #1 below. Your usage helps attract sponsors. If you
contact any of our sponsors, please be sure to thank them for
supporting BIOSCI. It is critical for them to get this feedback if
they are to continue their sponsorship for the long term.
Second, if you work for a company or organization that provides
products or services of interest to the biology community, please pass
this message on to your marketing or marketing communications
department or other appropriate group. Please ask them to help
support BIOSCI by sponsoring our Web site and explain the uses and
benefits of the system to the biology community. If they are
interested, they can then contact us for further information at our
tech support address, biosci-help@net.bio.net.
1) Using the WWW to access the BIOSCI/bionet newsgroups.
--------------------------------------------------------
As of 10 December 1995, all BIOSCI/bionet full newsgroups are
accessible through the World Wide Web (WWW) at URL http://www.bio.net.
One can read and reply publicly or privately to both recent postings
and archived messages through one's Web browser if it is configured
properly to send e-mail. Each newsgroup is equipped with its own WAIS
index. The main BIOSCI home page also has access to the BIO-JOURNALS
Table of Contents database WAIS index and the BIOSCI user address
database described in another item further below.
2) What to do about "spams," i.e., junk mail, ads, etc.
-------------------------------------------------------
BIOSCI is a set of parallel USENET newsgroups (the "bionet" groups),
mailing lists, and a hypermail archive at URL http://www.bio.net/.
The same postings are distributed on all media (except for a small
number of mailing-list-only groups at net.bio.net). Unfortunately it
is becoming a despicable practice on the Internet (by a few people out
to make a fast buck) to do automated mass postings to thousands of
newsgroups and mailing lists. These attempts to grab free advertising
are refered to as "spams" in the usual, somewhat boneheaded, net
terminology. USENET is more susceptible to this practice, and many
spams originate on the USENET groups and then are passed on to the
mailing lists. However, spammers also get lists of mailing addresses
and hit these too, so neither medium is immune.
What should you do personally if you get junk mail?
---------------------------------------------------
Just delete it and move on without reading it further. Filing a
protest is becoming increasingly useless because spammers are often
disguising the addresses where the messages are sent from. Unless you
really understand Internet mail systems, your attempt at protest by
sending replies to the message will often end up being sent to the
address of an innocent person that the spammer is victimizing.
What can BIOSCI/bionet do to protect its newsgroups?
----------------------------------------------------
The only solution currently available is to moderate the newsgroup.
If this newsgroup is already moderated, then you are in good shape.
Moderation protects the USENET distribution from about 95% of the
spams that are being sent to date and protects the mailing lists
completely. Moderation means, however, that someone has to take the
time to review each message before it goes out. We have set up
software here that simply allows the moderator to forward to an
address at net.bio.net messages that (s)he wishes to have distributed.
This takes no more time than that needed to read the message and pass
it on, say about 1 min. per message.
Most newsgroups currently have a discussion leader who is responsible
for their newsgroup. The discussions leaders and their e-mail
addresses are listed in the BIOSCI Information Sheet which is
available on the Web at http://www.bio.net/. If a newsgroup is being
hit with too many junk postings, please contact the discussion leader
for that group and see if there is interest in moderating the group.
Please do not assume that by simply posting a complaint to the
newsgroup itself, anyone on the BIOSCI staff will act on your
complaint. With close to 100 newsgroups to run, the BIOSCI staff has
to rely on the discussion leaders of each newsgroup to report problems
directly to us at biosci-help@net.bio.net.
We will moderate any of our newsgroups if the discussion leader tells
us that the readership of the group wishes to do so and if a moderator
is willing to do the work. For most BIOSCI/bionet groups, this
entails only a few minutes of work each day.
Moderating a newsgroup will resolve probably 95% of the junk postings
on the USENET distribution. Unfortunately there are easy ways for
determined spammers to override the moderation mechanism on USENET,
but we can protect our e-mail subscribers from unwanted postings if
the newsgroup is moderated. You can also access our newsgroups over
the WWW at URL http://www.bio.net. While this Web interface will not
stop spammers from trying to post to the groups, this will give you
yet another way, besides using USENET news, to keep the junk out of
your personal mail files. For those of you with local USENET news
systems, the Web interface will also give you faster access to new
newsgroups and recent postings.
3) Examples of subscribing and unsubscribing to the mailing lists.
------------------------------------------------------------------
PLEASE NOTE: The BIOSCI management does NOT act on
subscription/unsubscription requests that are posted improperly to the
newsgroups and mailing lists. People who do this only bother everyone
on the lists to no avail. Please be sure to follow the proper
procedures below.
Gory details are in the BIOSCI Information sheets on the Web at
http://www.bio.net. Below we give an example utilizing the
METHODS-AND-REAGENTS list at both of our two BIOSCI sites:
Users in the Americas and Pacific Rim countries who use the BIOSCI
------------------------------------------------------------------
node at computer net.bio.net:
----------------------------
A) Determine the "listname" which is the <=8 character mail address
^^^^^^^^^^^^^
for the group. These can be found in the BIOSCI Info. Sheet. For
the METHODS-AND-REAGENTS group the mailing address is
methods@net.bio.net. The listname is the portion of the address to
the left of the @ sign, i.e., "methods". The listname is used with
the "subscribe" and "unsubscribe" commands illustrated below.
B) Mail all commands in the body of a mail message addressed to
biosci-server@net.bio.net. Do NOT send commands to the newsgroup
posting addresses! Leave the Subject: line blank, any text on it
will be ignored.
C) In the body of your message put one or more of the following
commands with an "end" command on the last line, e.g.,
subscribe methods
unsubscribe methods
end
Do NOT put your e-mail address or other text on these lines. The
server only allows you to cancel your subscription if the address
on your mail header matches the address on our mailing list.
Please ask for help at biosci-help@net.bio.net if your address has
changed, e.g., if you know you are on the list but the server tells
you that you are not a member.
Users in Europe, Africa, and Central Asia who use the BIOSCI node at
--------------------------------------------------------------------
computer daresbury.ac.uk (also known as dl.ac.uk):
-------------------------------------------------
To subscribe and unsubscribe to/from the BIOSCI lists, you need to
specify the full USENET newsgroup name with "bionet-news." prepended.
The USENET newsgroup names are listed in the BIOSCI Information sheet
on the Web at http://www.bio.net/. For the METHODS-AND-REAGENTS list
the USENET newsgroup name is bionet.molbio.methds-reagnts, thus the
appropriate commands are
sub bionet-news.bionet.molbio.methds-reagnts
unsub bionet-news.bionet.molbio.methds-reagnts
These commands are included in a message addressed to mxt@dl.ac.uk,
NOT to the newsgroup mailing addresses. As usual, include the text in
the body of the message as text on the Subject: line is ignored.
To unsubscribe from all the lists at the UK node, use
unsub bionet-news
Please note that if the address in the list is different than the one
in your mail message header, you will not be able to unsubscribe by
this method. If you have problems, please mail biosci@daresbury.ac.uk.
4) The BIOSCI user address and research interest directory.
-----------------------------------------------------------
Please take this opportunity to add your name, address, and research
interest information to the BIOSCI User Address Database if you have
not already done so.
You can fill out the address form directly through our Web page at URL
http://www.bio.net/adrform.html.
The address database is reindexed nightly for WWW access (the URL is
http://www.bio.net/). If you are not directly on the Internet but can
reach it by e-mail, please use our waismail server to access the user
directory. waismail use is described above. You can also request a
user address form by e-mail from biosci-help@net.bio.net.
Please check your database entry from time-to-time to see if your
address information is still up-to-date. Because of our limited
personnel resources, we ask that you resubmit a *complete* form to
revise your entry; we only replace complete entries and do not have
resources to edit old forms.
From owner-population-bio@net.bio.net Mon Jan 12 22:00:00 1998
Path: biosci!agate!ihnp4.ucsd.edu!sdd.hp.com!vixen.cso.uiuc.edu!news-peer.sprintlink.net!news-peer-east.sprintlink.net!news.sprintlink.net!Sprint!sunqbc.risq.qc.ca!news.uow.edu.au!metro!metro!unsw.edu.au!not-for-mail
From: Russell Standish
and
FINAL INVITED PROGRAM
SECOND=20
INTERNATIONAL CONFERENCE ON COGNITIVE AND NEURAL SYSTEMS
May 27-30, =
1998=20
Sponsored by Boston University's
Center for Adaptive =
Systems
and=20
Department of Cognitive and Neural Systems
with financial support =
from=20
DARPA and ONR
How Does the Brain Control =
Behavior?
How Can=20
Technology Emulate Biological Intelligence?
The conference will =
include=20
invited lectures and contributed lectures
and posters by experts on =
the=20
biology and technology of how the brain
and other intelligent systems =
adapt=20
to a changing world. The
conference is aimed at researchers and =
students of=20
computational
neuroscience, connectionist cognitive science, =
artificial=20
neural
networks, neuromorphic engineering, and artificial=20
intelligence.
A single oral or poster session enables all =
presented work=20
to be
highly visible.
Abstract submissions encourage =
submissions of=20
the latest results.
Costs are kept at a minimum without =
compromising the=20
quality of
meeting handouts and social events.
Although =
Memorial Day=20
falls on Saturday, May 30, it is observed on
Monday, May 25,=20
1998.
CONFIRMED INVITED =
SPEAKERS
TUTORIALS
Wednesday,=20
May 27, 1998:
Larry Abbott, Short-term synaptic plasticity:=20
Mathematical
description and computational =
function
George=20
Cybenko, Understanding Q-learning and other adaptive
=
learning=20
methods
Ennio Mingolla, Neural models of biological vision
Alex=20
Pentland, Visual recognition of people and their behavior
Each =
tutorial=20
is 90 minutes long.
KEYNOTE SPEAKERS
Stephen =
Grossberg,=20
Adaptive resonance theory: From biology to technology
Ken Nakayama,=20
Psychological studies of visual attention
INVITED=20
SPEAKERS
Thursday, May 28, 1998:
Azriel Rosenfeld, =
Understanding=20
object motion
Takeo Kanade, Computational sensors: Further progress=20
Tomaso Poggio, Sparse representations for learning
Gail =
Carpenter,=20
Applications of ART neural networks
Rodney Brooks, Experiments in=20
developmental models for a neurally
controlled humanoid=20
robot
Lee Feldkamp, Recurrent networks: Promise and =
practice
Friday,=20
May 29, 1998:
J. Anthony Movshon, Contrast gain control in the =
visual=20
cortex
Hugh Wilson, Global processes at intermediate levels of form =
vision=20
Mel Goodale, Biological teleassistance: Perception and action=20
in
the human visual system
Ken Stevens, The =
categorical=20
representation of speech and its
traces in acoustics and =
articulation
Carol Fowler, Production-perception links in speech =
Frank=20
Guenther, A theoretical framework for speech acquisition
=
and=20
production
Saturday, May 30, 1998:
Howard Eichenbaum, The =
hippocampus and mechanisms of declarative memory
Earl Miller, Neural =
mechanisms for working memory and cognition
Bruce McNaughton, =
Neuronal=20
population dynamics and the
interpretation of dreams =
Richard=20
Thompson, The cerebellar circuitry essential for classical =
=20
conditioning of discrete behavioral responses
Daniel Bullock, =
Cortical=20
control of arm movements
Andrew Barto, Reinforcement learning =
applied to=20
large-scale
dynamic optimization problems =20
There will be contributed oral and poster sessions on each =
day=20
of the conference.
CALL FOR =
ABSTRACTS
Contributors are=20
requested to list a first and second choice from
among the topics =
below in=20
their cover letter, and to say whether it is
biological (B) or =
technological=20
(T) work, when they submit their
abstract, as described =
below.
*=20
vision * spatial mapping and navigation =
*=20
object recognition * neural circuit models =
*=20
image understanding * neural system models =
*=20
audition * mathematics of neural systems =
*=20
speech and language * robotics =
*=20
unsupervised learning * hybrid systems =
(fuzzy,=20
evolutionary, digital)
* supervised=20
learning * neuromorphic VLSI =
*=20
reinforcement and emotion * industrial applications
* =
sensory-motor=20
control * other
* cognition, planning, =
and=20
attention
Example: first choice: vision (T); =
second=20
choice: neural system models (B).
Contributed Abstracts must =
be=20
received, in English, by January 31,
1998. Notification of acceptance =
will be=20
given by February 28, 1998.
A meeting registration fee of $45 for =
regular=20
attendees and $30 for
students must accompany each Abstract. See =
Registration=20
Information
for details. The fee will be returned if the Abstract is =
not=20
accepted
for presentation and publication in the meeting proceedings. =
Registration
fees of accepted abstracts will be returned on request =
only=20
until
April 15, 1998.
Each Abstract should fit on one =
8.5" x=20
11" white page with 1" margins
on all sides, single-column =
format,=20
single-spaced, Times Roman or
similar font of 10 points or larger, =
printed on=20
one side of the page
only. Fax submissions will not be accepted. =
Abstract=20
title, author
name(s), affiliation(s), mailing, and email address(es) =
should=20
begin
each Abstract. An accompanying cover letter should include: =
Full=20
title
of Abstract; corresponding author and presenting author name,=20
address,
telephone, fax, and email address; and preference for oral =
or=20
poster
presentation. (Talks will be 15 minutes long. Posters will be =
up for=20
a
full day. Overhead, slide, and VCR facilities will be available=20
for
talks.) Abstracts which do not meet these requirements or =
which=20
are
submitted with insufficient funds will be returned. The original =
and=20
3
copies of each Abstract should be sent to: Cynthia Bradford,=20
Boston
University, Department of Cognitive and Neural Systems, 677=20
Beacon
Street, Boston, MA 02215.
REGISTRATION INFORMATION: =
Early=20
registration is recommended. To
register, please fill out the=20
registration form below. Student
registrations must be =
accompanied by a=20
letter of verification from a
department chairperson or =
faculty/research=20
advisor. If accompanied by
an Abstract or if paying by check, mail to =
the=20
address above. If
paying by credit card, mail as above, or fax =
to (617)=20
353-7755, or
email to cindy@cns.bu.edu.=20
The registration fee will help to pay for a
reception, 6 coffee =
breaks, and=20
the meeting proceedings.
STUDENT FELLOWSHIPS: Fellowships for PhD =
candidates and postdoctoral
fellows are available to cover meeting =
travel and=20
living costs. The
deadline to apply for fellowship support is January =
31,=20
1998. Applicants
will be notified by February 28, 1998. Each =
application=20
should include
the applicant's CV, including name; mailing address; =
email=20
address;
current student status; faculty or PhD research advisor's =
name,=20
address,
and email address; relevant courses and other educational =
data; and=20
a
list of research articles. A letter from the listed faculty or PhD =
advisor=20
on official institutional stationery should accompany the =
application and=20
summarize how the candidate may benefit from the meeting. Students =
who also=20
submit an Abstract need to include the registration fee with=20
their
Abstract. Reimbursement checks will be distributed after the=20
meeting.
REGISTRATION FORM
Second International =
Conference on=20
Cognitive and Neural Systems
Department of Cognitive and Neural=20
Systems
Boston University
677 Beacon Street
Boston, =
Massachusetts=20
02215
Tutorials: May 27, 1998
Meeting: May 28-30, 1998
FAX: =
(617)=20
353-7755
(Please Type or Print)
Mr/Ms/Dr/Prof:=20
_____________________________________________________
Name:=20
______________________________________________________________
Aff=
iliation:=20
_______________________________________________________
Address:=20
___________________________________________________________
City, =
State,=20
Postal Code: __________________________________________
Phone and =
Fax:=20
_____________________________________________________
Email:=20
_____________________________________________________________
=
The=20
conference registration fee includes the meeting program,
reception, =
two=20
coffee breaks each day, and meeting proceedings.
The tutorial =
registration=20
fee includes tutorial notes and two
coffee breaks.
CHECK=20
ONE:
( ) $70 Conference plus Tutorial (Regular) =
( =20
) $30 Conference Only (Student)
( ) $45 Conference =
plus=20
Tutorial (Student)
( ) $25 Tutorial Only (Regular) =
( ) $45 Conference Only (Regular)
( ) =
$15=20
Tutorial Only (Student)
METHOD OF PAYMENT (please fax or=20
mail):
[ ] Enclosed is a check made payable to =
"Boston=20
University".
Checks must be made =
payable=20
in US dollars and issued by
a US=20
correspondent bank. Each registrant is responsible=20
for any and all bank=20
charges.
[ ] I wish to pay my fees by credit card=20
(MasterCard, Visa, or Discover Card=20
only).
Name as it appears on the card:=20
_____________________________________
Type of card:=20
_______________________________________________________
Account =
number:=20
_____________________________________________________
Expiration =
date:=20
____________________________________________________
Signature:=20
__________________________________________________________