Well, I didn't expect much reponse since the other posts about this subject received no reponses. Anyway, he is the approach I took. I just put the needed...
I probably should have elaborated on MakeDecision(). public void MakeDecision(DataRow dr) { // Clear the network box.ResetState(); // Reset all inputs. ...
The but from Item 3 below, which forces at least 2 outputs to be defined, appears to be in SharpNeat.Decoders.HyperNeat.Substrate, line 299: // Read bias...
Hi, ... The code you're referring to takes the evolved CPPN and queries it to create the neural net that you will go on to evaluate. The CPPN accepts inputs...
I set the outputs to 1 in my expermiment class: public int OutputCount { //get { return 1; } get { return 2; } } When I return 1, I get the error I described....
... So I would set OutputCount to 1, and also modify Substrate.cs so that it doesn't try to read the bias weight from output[1]. Basically you can comment out...
I am testing NEAT against the problem of detecting fraud. My approach is different than many of the examples because it is not a robot/game approach where you...
As additional information to my stated problem, the best genome and connections I show are from generation 3821. Also, the genomes do complex, but then...
Major bug in my code when setting input array values: for (Int32 j = 0; j < box.InputCount; j++) { box.InputSignalArray[0] = Convert.ToDouble(dr[j]); } should...
Perhaps I'm far too tired to think this through clearly. I was thinking of adding a "random search" implementation to my NEAT experiment for simple comparison...
Hi Drew, I would lean towards assigning random fitness values. If you assign everything an identical fitness, then in some NEAT implementation it is likely...
Ken, Thanks for the input. I'll change it out to random values when I start running my experimental conditions "for real" (still fixing data collection). I...
This is really a novices question, but having just started using HyperNEAT (as opposed to plain NEAT), I find that with the weight constaints used in the...
http://www.youtube.com/watch?v=KMy2zRN4d38 I hope to check the code into SVN soon, there are also a few fixes that have been submitted to the SVN repository...
... Hi Colin, Have you considered using novelty search for the Bipedal Walker? (From Ken Stanley's "Abandoning objectives: evolution through the search for...
There may be some underlying issues in Colin's biped domain that I'm not aware of, but from my experience so far I think this would be a great domain for...
Hi Rob, you may indeed be experiencing weight saturation from having so many inputs. (Which version of HyperNEAT is it, by the way?) One idea with large...
Hi Ken, Thanks for the suggestion. It was indeed correct. I've scaled the input values, which were typically close to or at +/- 1.0. I used the number of input...
Hi all, I've just started a Ph.D. with a topic around neuroevolution, online-learning, adaptive NNs (via synaptic plasticity and/or similar). In my feverish...
Hi Oliver, that's great that you were able to start a Ph.D. on this topic. Where are you doing the Ph.D.? I look forward to seeing your ideas on evolving...
Hi Ken, Thanks for your reply. I'm at the University of New South Wales, Australia (same place I did my Honours project studying properties of HyperNEAT, with ...
Hi Oliver, I don't think it hurts to point to your blog on here if you use your judgment regarding the relevance of the posts. Even if they are not explicitly...
Just wanted to give a heads up that I've finished my experiments and analysis of using NCD for novelty search in competitive coevolution, and even with my very...
SharpNEAT V 2.2 as a 7zip archive of source and precompiled binaries is available at: https://sourceforge.net/projects/sharpneat/files/sharpneat2/ From the...
I am puzzled by the result of band pruning in Figure 3: Example point selection in two dimensions, from Evolving the Placement and Density of Neurons in the...
Hi Tony, I believe the issue there is with the definition of a "band": "points that are enclosed by at least two neighbors on opposite sides (e.g. left and...
What I mean is the band pruning seems too restrictive. Suppose a CPPN substrate is being pruned for a Go game board evaluator. If ideally the weights along...
