From: Sebastian Smith <[EMAIL PROTECTED]>
To: Brandon Mitchell <[EMAIL PROTECTED]>
CC: [email protected]
Subject: Re: [RLUG] Hello RLUG
Date: Fri, 1 Dec 2006 23:48:58 -0800 (PST)
I didn't really describe what I've been working on in my original email, so
I'll contribute that to this thread to keep it rolling.
well no... because we all already knew about the accident with the
monkies...
For the past few months I've been working on the design of a recurrent
neural network for use in signaling formation changes in groups of robots.
The project is focused on machine learning, but I've naively tied it to
robotics so it appears to be related to my research :) What I specifically
want to do is generate neural networks with a genetic algorithm, and
evaluate their performance by looking directly at the structure of the
neural network rather than its output values.
what do you mean by "formation changes?"
why would you want to evaluate the structure of the network rather than the
output values, isn't the whole point to use the GA to generate some
structure that's too complicated/non-obvious to have preconveived that
outputs the desired solution? what's the point of having a fancy network
that doesn't produce anything?
what tools are you using for your GP stuff? lisp? :-) are you reading koza's
books? :-)
did you see the earlier rlug thread on mosquito lisp[1]? it's a
concurrency-oriented lisp inspired by scheme and erlang...
[1] http://bc.tech.coop/blog/061119.html
I'm unsure of the benefit of this method (if
any) over simply training the outputs of a neural network.
then why are you doing it
I'm also
unsure if the network I'm working with should even be considered a neural
network since it's structure is slightly different.
who cares, it's interesting ;-)
Regardless, I'm experimenting... I can tally the results later. (This is a
class project BTW, it is not related to my research).
I'm also working on another neural network project in a different class.
This time I'm using them for function approximation - a common application.
I'm going to compare neural network performance to other approximation
techniques (empirically this is easy, but I'm not sure how to accomplish
this on the theoretical side... more reading I guess).
As far as research goes, I'm making groups of robots communicate reliably.
I'm integrating multi-hop routing algorithms with 802.11a/b/g equipment to
produce wireless meshes. I'm then then building a robust communication
mechanism based on the publish/subscribe paradigm on top of the meshes.
It's a big project, and will take quite a while to complete.
is this an exercise in design or implimentation?
hrm, well, cool... back to homework I guess...
peace
Nick
Guess that's most of what I've been doing... aside from socializing and
whatnot :) *mmm* beer.
Will the next meeting be at the new GBIS offices?
- Sebastian
On Fri, 1 Dec 2006, Brandon Mitchell wrote:
I'm still working on the SOA/REST/XML-RPC interop "stuff" in Ruby and
building data warehouses for feeds from various and sundry entities
around Northern Nevada. We're aggregating all kinds of strange data in
preparation for a major corporate reorg (ugh!) at which point, I'm
told, we'll need all this crap. Truth be told, not even Corporate
knows what the next 18 months will bring.
On the Linux front, I've finally got our new servers up and running
(it took ages to get them in). They're Dell 2950s with Woodcrest
Xeons, 8GB of RAM, 4 x 73GB SAS in ext3 (no XFS in RHEL!) RAID10
running 64-bit RHEL4. They are in a multi-master MySQL 5 (MySQL AB
binaries) setup across our two NOCs (there's a dedicated, redundant
10G fiber link between the two NOCs). So far, I'm getting ~1700
Transactions per second out of the setup on tables of between 8 and
16G in size (several million rows), which I think is pretty darn good
(for a Dell). All in all the experience has been pleasurable with
these Dells; it only took me about a day to get them installed from
bare metal and replicating. On the long game, RHEL gets a bit far
behind on patch level compared to other distros, but when it's only
running MySQL and I can get reliable updates directly from MySQL AB...
it's not so bad.
Anywho... PING back!
Brandon
On 11/29/06, Sebastian Smith <[EMAIL PROTECTED]> wrote:
Hey everybody,
Long time no speak. Where the heck is everyone? The list has been
pretty
quiet lately. It's crunch time at the Uni. As usual, my projects aren't
done yet... so I have two weeks of hell upcoming. Fortunately, it should
be nearing completion by the time the next meeting rolls around. That
means I'll be ready to absorb knowledge and alcohol... *hehe*.
I've no new news on the Linux front. I haven't done anything cool for
the
past few months. I guess I'm just a user now.
So... PING.
--
( )
H
H
_H_
.-'-.-'-.
/ \
| |
| .-------'._ Sebastian Smith
| / / '.' '. \ Robotic Research Laboratory 775.784.4580
| \ \ @ @ / / Department of Computer Science and Engineering
| '---------' University of Nevada, Reno
| _______| http://www.cse.unr.edu/~ssmith
| .'-+-+-+|
| '.-+-+-+| (Ascii art by Silver Saks)
| """""" |
'-.__ __.-'
"""
_______________________________________________
RLUG mailing list
[email protected]
http://lists.rlug.org/mailman/listinfo/rlug
--
If UNIX doesn't have the solution you have the wrong problem.
UNIX is simple, but it takes a genius to understand it's simplicity.
_______________________________________________
RLUG mailing list
[email protected]
http://lists.rlug.org/mailman/listinfo/rlug
_________________________________________________________________
MSN Shopping has everything on your holiday list. Get expert picks by style,
age, and price. Try it!
http://shopping.msn.com/content/shp/?ctId=8000,ptnrid=176,ptnrdata=200601&tcode=wlmtagline
_______________________________________________
RLUG mailing list
[email protected]
http://lists.rlug.org/mailman/listinfo/rlug
