EvBrain: A Software to Evolve Artificial Brains (V. 3.02 Release)
Chopin
artificial brain evolution. It uses a topographical developmental
method to compose the brain and
connections.
Web Page:
http://brainew.com/research/alife/EvBrain/
Publication:
Jung, Sung Young, "A Topographical Development Method of Neural
Networks for Artifical Brain Evolution", Artificial Life, The MIT
Press, (to appear).
Enjoy Evolving Artificial Brains.
Kent Paul Dolan
> A new version of EvBrain was released. EvBrain is
> a software for artificial brain evolution. It uses
> a topographical developmental method to compose
> the brain and connections.
Do you really mean "topographical" == "elevation map
oriented", or did you mean "topological" ==
"connection oriented"?
> Web Page:
> http://brainew.com/research/alife/EvBrain/
You need to replace the JPG images by higher
resolution GIF or PNG "lossless" images. The JPG
artifacts make the words unreadable, and the images
incomprehensible. Also, the image "The Nature
Field" is too low contrast (green on gray) to be
understood.
Still, you seem to have a promising approach.
xanthian.
--
Posted via Mailgate.ORG Server - http://www.Mailgate.ORG
Chopin
> "Chopin" <chopinx...@hotmail.com> wrote:
>
> > A new version of EvBrain was released. EvBrain is
> > a software for artificial brain evolution. It uses
> > a topographical developmental method to compose
> > the brain and connections.
>
> Do you really mean "topographical" == "elevation map
> oriented", or did you mean "topological" ==
> "connection oriented"?
>
"topographical" does not mean "elevation map". The word
"topographical" is a widedly used terminology in neuroscience. This
term is used to designate a certain type of inter-layer connections.
You can understand as following.
If two layer of neural connections is topographically connected, than
any two near neurons in a layer are connected into two near neurons in
another layer.
For easy understanding, just imagine parallel connections from one
layer into another layer with distortion.
> > Web Page:
> > http://brainew.com/research/alife/EvBrain/
>
> You need to replace the JPG images by higher
> resolution GIF or PNG "lossless" images. The JPG
> artifacts make the words unreadable, and the images
> incomprehensible. Also, the image "The Nature
> Field" is too low contrast (green on gray) to be
> understood.
>
I replaced the images.
> Still, you seem to have a promising approach.
>
> xanthian.
How to develop neural structures from a gene sqeuence is one of very
important issues in artificial life, and neuro-evolution.
Thank you for interest.
__
Sung Young Jung
http://brainew.com
Kent Paul Dolan
> "topographical" does not mean "elevation map".
To a map maker it does:
Topographical surveying, that branch of
surveying which involves the process of
ascertaining and representing upon a plane
surface the contour, physical features, etc.,
of any portion of the surface of the earth.
http://dict.die.net/surveying/
and I was in that business from 1973-1981.
The usage in Neural Nets is a bit dodgy, since there
isn't any spatial metric distance between neurons,
only a graph theoretic "edge weight" distance, but
I'm not about to try to revise a science in which I
don't participate.
> I replaced the images.
Thanks, those are much easier to see.
The words went missing from the one that formerly
had a control panel, though.
Phil Tomson
In article <2d6bfedf.04080...@posting.google.com>,
Will the source code be available for examination and so that we can
compile on Linux?
Phil
Chopin
> "Chopin" <chopinx...@hotmail.com> wrote:
>
> > "topographical" does not mean "elevation map".
>
> To a map maker it does:
>
> Topographical surveying, that branch of
> surveying which involves the process of
> ascertaining and representing upon a plane
> surface the contour, physical features, etc.,
> of any portion of the surface of the earth.
>
I know the dictionary definition. It may give confusion to people who
are not familar with this area. However, to prevent confusion, take in
mind "topographical connection" rather than "topographical map". The
main issue that I wanted to solve in the research was to create
topographical connections with low computation. "topographical map"
has no relation to my research.
>
> The words went missing from the one that formerly
> had a control panel, though.
>
I don't think the picture of control panel is important to be shown on
the website. You can see the control panel when you download and run
the EvBrain software.
Thanks.
__
Sung Young Jung.
http://brainew.com
Chopin
> In article <2d6bfedf.04080...@posting.google.com>,
> Will the source code be available for examination and so that we can
> compile on Linux?
>
> Phil
I am planning to open the source code under the GNU license after my
paper is publised on the journal (maybe in upcoming Autumn). However,
if somebody wants to port the source code into other platforms
including Linux, I will consider to provide the source code.
Thanks.
__
Sung Young Jung
http://brainew.com/
7
> A new version of EvBrain was released. EvBrain is a software for
> artificial brain evolution. It uses a topographical developmental
> method to compose the brain and
> connections.
>
> Web Page:
> http://brainew.com/research/alife/EvBrain/
>
> Download:
> http://brainew.com/e
ezboard.cgi?db=alife_neuroevolution&action=down&dbf=200408030000&downfile=EvBrain3.02.zip
>
> Publication:
> Jung, Sung Young, "A Topographical Development Method of Neural
> Networks for Artifical Brain Evolution", Artificial Life, The MIT
> Press, (to appear).
>
> Enjoy Evolving Artificial Brains.
Is this a sock puppet program?