Hi Everyone, I am pleased to announce the opening of Picbreeder, a massive experiment in collaborative interactive evolution, which happens to be based on...
This is the first public test of Picbreeder so of course we will have technical issues here and there as we ramp up. I hope you will let us know if you...
Yes, we're trying to optimize some of the way we distribute processes and it caused a few crashes- sorry about that. We do have sufficient computational...
Hi there, I am planning to create an application of CPPNs that generate 3D tetrahedral meshes to be used in physically-based simulations... The CPPN will...
Mridul, I think NEAT could be an interesting approach to the repeated auction, but I am not sure whether your testbed really simulates the dynamics of such a...
Thanks for the note, Ken I realize that passing just the previous reward is not a good input for learning. I changed the ANN configuration to have a single...
It depends on the type of application it is to be used. An exmple one might be evolving a structure that minimizes the stress at specified area, given an...
Hi, I have recently started using the Matlab NEAT and noticed what seems to be a duplicated connection from the bias to the output nodes. The configuration of...
Sorry! Solved already. The program starts with a vector of connected nodes, which must be commented out or modified if using a network configuration different ...
Hi, im new here and also new at NEAT. Im really interested to know more about the coevolution setup K. Stanley used in the paper "competitive coevolution...
... I cannot speak to Ken's setup; however, you could use something similar to chess tournaments. Check out the Glicko2 rating system and the Swiss Tournament...
Yes, I agree, you could get some quite interesting solutions from such an approach. ken ... one ... specified ... not ... 3D ... simulations... ... properties ...
I've never seen any paper about coevolution using glicko..but it does seem interesting. thanks. Well, I also hope someone can help me clarify about Ken's...
It is neither A nor B. The question is how the very first generation could proceed when there is no information about either population to begin with. Our...
I am looking for any suggestions on how to discretize the output from a CPPN into finite elements.. You know that a CPPN's output is actually smooth and has...
Peter, Given the computationally intensive nature of FEM based analysis, you might consider evolving agents which operate on a single element at a time in the...
This is pretty neat! ...both puns intended. :-p Do these things start from a picture (e.g. a car)? Congrats on pulling this off. Very cool. Cheers, Jeff Clune ...
Thanks Jeff, Picbreeder was a lot of work to put together! In answer the your question, remarkably, everything you see on the site started ultimately from...
By the way you can follow the ancestry of any image back to its first ancestor (i.e. the first image that someone evolved from scratch that is a link in the...
Hi Ken, all, Nice job on the picbreeder site BTW. Y'know I was just wondering what results you would get if you tried to evolve towards a given image, say by ...
See my responses below. ... I agree. That is simply incredible. A standing ovation is in order for you and all involved. Last night when looking at the...
... Like Jeff, I'd be interested in hearing you expand your thoughts on this. I think the demonstration is no less impressive, but it seems to me that the ...
Hi Colin, Please see my response to Derek and Jeff with regard to evolving to a target, since actually it's relevant to their question and yours. About the...
Derek and Jeff, I would be happy to expand on my controversial remarks on interactive evolution being more "natural" than the usual kind of evolutionary ...
Good points, Thanks for that :) A simplification (perhaps oversimplified) is that TBE: - limits the solution, putting a ceiling on it's possibilities -...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
