Hello, I am experimenting with my implementation of HyperNEAT in C++. The application is a roving eye visual discrimination system. The CPPNs use sigmoids,...
Hey Peter, Good to hear that you are making use of the software! As soon as I get through this next milestone, I'll release a 1.1 version which has support...
Jason or Ken, I am attempting to understand the nature of inputs & outputs as applied to both the Hypercube and substrate networks, without being limited to...
Yes, a 2D space was the first realization. But things got complicated mainly because there is no way to make a matrix-like input layer, and to place the hidden...
Yes I would suggest trying the experiment without a roving eye since in a sense the problem of dealing with different parts of the same field in the same way...
Andy, I'd say this is a about right. I elaborate a bit below each ... general ... This statement is correct although you might do things that you would avoid...
Petar, I have used a n-dimensional context space to hold the input and output nodes, and embed the hidden nodes (graph) - genome dual in that space. If you...
Ken, Out of curiosity, what have you been able to do using this representation? ken ... output ... space. ... 2D by ... the ... involving ... Of ... ...
Ken, One troubling "detail" is that of differences in dimensionality between substrate inputs and outputs. In the examples used in the various papers, the...
Andy, There are still a lot of ways to configure the substrate for such a problem, so it's difficult to give a general answer. For example, you might have a...
Ken, The cited example modification of the paper geometry problem was only to provide an instance in which the dimensionality of the substrate inputs and...
Andy, Sure, I understand it was just an example. I went with the example just for illustration, but let me give a shot at a more general answer: I think the...
Ken, Two issues come to mind. The first is degree of required user intervention in the determination of substrate topology. The second issue is embedability. ...
Andy, I think all these ideas are valid and deserve to be explored. Automation is a potentially significant future direction so heuristics for it are an...
Ken, I have a paper being submitted to the Software and Systems Modeling Journal (SoSyM) that illustrates some accomplishments. If it is accepted, I will be ...
Ken, Recursive Supervisor NEAT is reformulated as follows: (a) NEAT moves from namespace to supervisor class; (b) namespace is removed and all classes folded...
... So I think your newest statement of HyperNEAT's advantages is fine, and basically I'm with you until about here. Basically I don't see this discussion as...
Andy, I see, I understand the rationale. I will be interested to see how things go with your ideas. The main question will be whether the system crosses the...
Hi, a beginners question on understanding fitness sharing. I've referrd to http://tech.groups.yahoo.com/group/neat/message/2292 which gives an outline of...
Hi, The original fitness is simply the fitness of the individual you assign based ot your experiment :) You divide this fitness by the number of individuals in...
Hi, I downloaded rtNEAT and there was no NEAT configuration file for single-pole experiment (.ne file). If anyone has a decent version of the *.ne files (the...
Petar, Thanks for the prompt response. Think I understand it now. Also since found Ken's Phd thesis has it on page 43. Adjusted fitness looks like a weighted...
... Yes, this is the usual way to do fitness sharing. You can notice (in the NEAT FAQ on Ken's home page) that there is also another way to try. You can use...
Ken, While your point is well taken, the significance of the infrastructure differences lies not primarily in the diversity of building blocks, with possible...
Andy, It will be interesting to see whether taking a prior evolved network and designating it as the function of a single node in a higher- level network will...
Ken, The common underlying thread is that unless adequate infrastructure in in place, there is no way to even attempt anything other than that which fits...
Given an arbitrary point "P" in design space, comprised of the following: (a) problem specific inputs (PSIs); (b) auxiliary additional parameters associated w/...
It would seem that the issue of substrate hidden nodes is mitigated somewhat by the fact that in certain instances they are nothing more than duplicated...
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.
