Hi, We're two students and our task is to find a NN that is able to control a hexapod robot. To find a suitable NN, we want to use a Genetic Algorithm. Inputs...
Hey guys It is possible, I am just porting this bit of functionality from my original AI framework. Basically, you have to create a class that extends ...
Hi, I'm trying to do something like that but with unreal tournament bots. Instead of overriding the initialise method I extended the NEATFitnessFunction where...
Hi Cícero There's nothing wrong with that approach - infact seems a very good idea. One thing to be aware of though, is there is only one instance of the...
Hi Matt, I have knowledge of these circumstances, and I'm afraid of my approach take too much time to evolve the NNs. That is why I want to use multiple...
Matt, I'm already using the approach I metioned on my first email, but I'm interested in using the distributed learning of neat4j. All I have to do is to pass...
Matt, I'm using a game mod called GameBots that let me connect a bot to a unreal tournament server game, so the bot receive sensorial information from the ...
Any class that passes between the server and the remote agent must implement the serializable interface (and any class contained within etc). I suspect yuou...
... Ok. I was thinking that way, but even with my classes implementing the Serializable interface I still got the error. Maybe I'm just doing something wrong. ...
Hi, I am new to this site, and I am very interested in doing reearch related to the NEAT framework, but I am having som problem compiling the release. First, I...
Leo, You might try looking into the "FILES" area of the group website for another source closer to your platform/compiler. Good Luck, Andy ... rtneat. ... C++ ...
Hi, If we initialize NEAT with any population size, 28 input neurons, 18 output neurons and config parameter FEATURE.SELECTION=false, we get a null pointer...
Hi again, We just looked through the relevant code and most probably found a bug within the class InnovationDatabase.java. Here the code from line 90 to 116 ...
Iwas able to compile rtNEAT :-D, it was just a gcc problem, just set it to compile with gcc3.4 and good .. however for NEAT, even with the patch and compiling...
Unfotunatly I am not sure how easy it is to get gtk and guile working on new machines/platforms and I'm also interested in others' experiences with this....
Matt, There is the base code that I'm using in my experiments If you could take a look I would be glad Best regards, ... -- Cícero Augusto - Bacharelando em...
Hi I can't get your attachments. It would be better if you mail me them direct. Cheers Matt ... your ... me the ... times ... distributed by ... take ... to a...
Colin, I apologize if this is something that was discussed earlier, but could you comment on what makes SharpNEAT so fast? I was playing with the latest...
... Hi Ken, sure no problem. It's easiest if I explain from the beginning so apologies if some of the explanation seems obvious and/or long winded... One of...
Hi, I would like to know what kind of learning the ANN's in NEAT have or it is a new kind of learning? Sorry for the dumb question []s -- Cícero Augusto -...
Hi I'm taking a look for possibly using in for my Masters project, which needs evolving the architecture for a neural net for a real world application. I have...
Sorry , just to clarify for (3) if we took an example with say 300 training samples, with say 100 samples in the test set. Just to try and get an idea. Mike...
In most experiments, the individual ANNs in NEAT do not learn. Instead, learning happens over the course of generations through evolution. In other words,...
... But this learning over the course of generations through evolution can be classified as some "standard" kind of learning? ... -- Cícero Augusto -...
I know people have combined NEAT with BP in the past. For example, Chris Christenson had some success with it, but I'm not sure he's still reading this group....
Ken, Thanks for your answer, I was thinking in use the terms supervised, unsupervised or reinforcement learning but none of them can be applied to this case...
There are also some papers by Nolfi, Parisi, Floreano and the like (as well as others) that apply back propagation to the networks during trials and evolve...
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.
