Hi all, Firstly, I apologise in advance for the "newbie-ness" of the questions I'm going to ask... I see that Ken's code uses a floating point sigmoid...
Hi everyone, I'm new to the NEAT group, so let me introduce myself shortly: I live in Hungary and work as a software developer (currently at Lufthansa ...
Hi, I found only one or two threads about testing of NEAT implementations. I think that extensive unit testing is one of the best ways to ensure code quality,...
Hi, I've seen that there were a lot of discussion about network visualisation in the past. However, the goal of most of them was to generate the image of the ...
Hi, I was playing with the idea of a kind of structure sharing between networks and I'd like to hear your opinion about it. Problem: algorithms related to the...
Hello, I use three approaches: 1) I run several tests with a pseudo-random-generator using known seeds. This produces a pseudo-random sequence of numbers...
... I'm not sure if Ken addresses that directly...can't remember. Additionally I was just wondering is there any benefit from using ... Yep, I've heard this...
... Well, when looking for ways to visualize data and networks, we looked for existing opensource tools, and basically didn't find any we liked. In ...
... Correct me if I'm wrong, but doesn't magnitude of weights make up for slope of activation function? Smaller weights cause the output values to fall near...
Hi Derek ... I agree that a nice opensource solution would be the nicest... I found a java app that can do exactly this, and does it very well. There ... Derek...
For ANJI, we do some tests like Robert's #1 below. The only problem with this is it makes some tests very brittle. Often the tests will break just because...
Hi Jim, On 2/6/06, Jim O'Flaherty, Jr. <jim_oflaherty_jr@...> wrote: Sandor, ... I've already known about Semiann, but I thought its killer feature is...
Sandor, Thank you for your acknowledgment. ... mean do you know how much time your optimizations can save (at least for some specific experiments)? What would...
Hi All, I want to buid an nnPlayer to play a board game. Could anyone suggest kindly how to do this as I am very new to ANJI and an detailed explanation in how...
... Hi Ramin, This isn't ANJI-specific, but have you read our paper on evolving an NN for Go? It's at: http://nn.cs.utexas.edu/keyword?stanley:gecco04 That...
Hi ken, Thank you for your kind email, I didn't realize that they had Tic Tack Toe example, anyway this will be great help to me as I want to build topologies...
I assume you've been to the website, http://anji.sourceforge.net. I would recommend: 1) download the distribution http://anji.sourceforge.net/downloads.htm 2)...
Ramin, My only advice is that it will be a challenge to start learning directly on a 12x12 board. That's a pretty big area. However, it will be interesting...
Hi, OK so this is me just scanning a log-file while the bath fills (having left it SLEAT running while I watched CSI - what-do-you-know, the daughter did...
Hi Everyone, I've spoken to Derek James (who runs this group) about this issue, and he suggested I bring it up for discussion. So please feel free to share...
One other possibility is that in December people went on vacation (so to speak), and nobody has really returned yet... Thus, nobody has said anything to spark...
Ken, Help me out. I am not sure what you are referring to when you talk about the Archives being open to the public? Do you mean the Yahoo!Group posting...
Sorry about that Jim. What I am referring to are simply the messages and nothing else (for example this post is a message). I used the word "archives"...
Hi, Those of us who read the messages by mail have to be subscribed anyway... How does the activity compare to the same period in previous years... 'cos you do...
... I guess that included me... ;-) Last November I sparked a discussion about speciation thresholds, but I haven't been an very active participant since then....
... Yeah, if they're really low-level implementation details, no need to post here. If it's algorithmic, then it's probably a good idea (since someone may be...
I'm generally in favor of leaving the archives open, but it matters how you're using the list. If you leave it open the list roster would have folks who want...