Re: Greetings all, new kid here :-)

Funny that you bring up NEAT. I'm currently working with Ken Stanley on
integrating NEAT into a program of a very large scale. Its quite an
enjoyable project.

<[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]
> RobbieGee wrote:
> >
> > Okay, done reading the tutorial, thanks for the link =) Actually there
> > wasn't much new to me, but it was well written and cleared up a couple
> > things for me. Mission accomplished ;-)
> >
> > How much is actually known about NNs? I'm having difficulties mapping
out
> > what's old and tried versus what is "today's topic". Using GA, do you
> > actually evolve entire NNs, or do you just adjust the weights?
>
> A great deal is known about NNs, but it's a tiny fraction of what might
> be known.  I don't personally know very much, but, like you, I'm
> learning.
>
> If you want to know about a very different type of NN, download the
> 4play package.  You can read the release notes first if you're not sure
> that you want it.
>
> I use GA just to adjust the weights, except sometimes I have adjusted
> the number of neurons also.  (That's in a package called annevolve)
>
> There is some software in C++ that evolves the entire topology.  It's by
> Stanley and Miikkulainen.   It's called NEAT, and you can get it here:
> http://nn.cs.utexas.edu/soft-list.php
>
> m
>
> -- 
> "Many are stubborn in pursuit of the path they have chosen, few in
> pursuit of the goal." - Friedrich Nietzsche
>
> http://annevolve.sourceforge.net is what I'm into nowadays.
> Humans may write to me at this address: zenguy at shaw dot ca