Jump to content

Google Frog

Member Since 30 Jun 2007
Offline Last Active Sep 14 2008 12:02 AM

Posts I've Made

In Topic: Creating Ai That Learns

02 March 2008 - 11:44 AM

Can we revisit This Post.

Posted Image

Have you made this NN yet? If not I would like to have a go. It seems you've come up with 2 ideas here; self training and stacking, I'll start with self training which is the net up the top left.

I have a few questions about the net.

The light red layer is the logical output. Or...the output the network thinks it's action will produce.

By output did you mean the same as action? Shouldn't the net already know what action it is doing.

Last, but not least, the yellow dot is what the AI predicts the training variable to be

Where does this variable come from? Is it based on external stimuli?

In Topic: Gmphysics - Version 5 Prebeta Released

05 February 2008 - 08:25 AM

Is there a way to save/load polygons to text files? If not this would be useful for V5 because create_polygon_auto takes a long time with large sprite so calculating the polygon only once would reduce loadtimes.

In Topic: Gmphysics - Version 5 Prebeta Released

30 January 2008 - 10:26 AM

Not sure what you mean by laser body collisions; if you mean raycasting

I think these a bug with get_ray_body()

In Topic: Gmphysics - Version 5 Prebeta Released

28 January 2008 - 09:02 AM

In version 5 will laser body collisions be fixed? Also it would be nice if reluctant bodies could still attract other bodies with magnetism but not move themselves.

In Topic: Creating Ai That Learns

28 January 2008 - 06:39 AM

I like the idea of a competition they usually create lots of discussion. I'll probably join but you need to define it more. I would like to know what inputs the AI will have and what it will be able to do.