Hey Colin, Or anyone else who is interested. Does anyone have any recommendations for papers on intraday financial time series analysis? I have quite a library...
I'm a Mechanical Engineer and I'm rather new to Neural Networks. I'm currently doing research in Q-Learning in the area of walking robots. I ran across NEAT...
Maybe the best thign to do would be to contact Shimon Whiteson and find out if he's able to share his own NEAT+Q code. He is the one who created NEAT+Q and he...
Thanks, I'd really appreciate it. I'll see what I can do about getting in touch with him. ... a ... space ... 2 ... link ... clear ... position, ... 10 ... ...
... I'm new to Q learning (since reading your post here a few days ago). I've read about it a bit, and wonder how useful it is for thing other than state...
Hello, This email message is a notification to let you know that a file has been uploaded to the Files area of the neat group. File : /TSNEAT Source.zip...
neat@yahoogroups.com
Apr 10, 2007 11:00 am
3101
Is there an installation how-to anywhere? Can't seem to run make on it - do I need cmake or something to compile these files? Cheers Tom ... -- PS Note this...
OK I was too lazy for now to properly set up the Source Forge project. I've put a copy of my source code in the files section. The project should compile quite...
Hey Tom, Sorry I usually import it into KDevelop and compile it from there. I will try and get it to compile via command line make post the updates here soon. ...
OK I will fix this properly in the next version. But to make the two executables this is the workaround for Linux. cd to the "neatl" directory and create a...
I have just finished posting my entry into the NEAT world: https://sourceforge.net/projects/aipac AIPac (or PacAI - I tend to swap them) is a program that...
Kirk, this program looks really cool but it crashes for me right after it starts up. It manages to show the initial screen (I see Pac Man on the board) but...
... It did run for me (Windows XP, P4@3GHz, 3.25GB RAM). I didn't know how to control it the program though. I did watch the video too. I'm wondering what...
Wouldn't you know it the one that crashes is Ken's?? Craziness. There are no dependencies other than the various files in the Resources directory and the...
... re: Inputs.. I wonder if some global information in addition to your local observations may be useful? For example, you may divide the grid up into 2x2...
So a few days back I mentioned some adaptive representations work I've been doing using NEAT to study the effects of "development" on evolvability. This...
I'm disappointed there are crash issues cropping up. I'll do some tests and make sure I haven't missed something in the distribution. I like your idea for...
Here's a couple of things to try: In the parameter file AIPac.param, find the Replicates parameter which should be 10. Set this to 2. You can also make the...
Hey everyone! I'm pleased to announce that I have a software release of my HyperNEAT package that I used for the GECCO 2007 paper. The software is released...
I am sure this must have been done but I can't find any papers on it. Has anyone tried starting out with an initial population that connects all the inputs to...
Hi Shane, I don't know what the identity function is exactly.. I need to learn that. But I suppose you can also try to split the neuron activation into stages,...
Identity function is f(x) = x, simple =) I've thought about multi-activation-functions populations as you 2 proposed. My contribution here is to add the sine...
This is the Identity function? Heh, I call it linear.. So I have this already :) You're right that a simple mutation will have a significant impact, but what...
Totally agreed! Your solution seems good for the problem. About the 4 functions (MIN, MAX, ADD (default), MUL), do you mean apply them before the activation...
Well, in some point of view, Neuro-Evolution is a genetic programming scheme, composing many many "1/(1+exp(-(weighted_inputs * x)))" in such a way, that the...
... I didn't run it (I'd love an OS X binary), but I took at look at Jason's code for a few minutes last night. It is one of the cleaner NEAT implementations....
Well I am not sure if I would personally add in specific operators like add multiply, std dev etc. This makes the problem start to stray towards Genetic...
I haven't tried to compile my application for OS/X, but I think if you follow the instructions for linux, it should compile fine. I can try to create an OS/X...
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.
