Gneural Network - Tasks: task #14206, implement HyperNEAT for genetic...
You are not allowed to post comments on this tracker with your current authentication level.
task #14206: implement HyperNEAT for genetic algorithm training.
Submitter: | Ray Dillinger <rayd> | ||
Submitted: | Sun 30 Oct 2016 05:58:37 PM UTC | ||
Should Start On: | Sun 30 Oct 2016 07:00:00 AM UTC | Should be Finished on: | Tue 30 May 2017 07:00:00 AM UTC |
Category: | None | Priority: | 5 - Normal |
Status: | None | Privacy: | Public |
Assigned to: | None | Percent Complete: | 0% |
Open/Closed: | Open | Effort: | 0.00 |
No files currently attached
Depends on the following items: None found
Items that depend on this one: None found
Carbon-Copy List
There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.
No changes have been made to this item
Powered by Savane 3.13-02a9.
Corresponding source code
HyperNEAT is the same general algorithm as NEAT in terms of tracking the point of origin of given genes as a guide to how genomes can be combined.
But rather than evolve a network directly, it evolves a set of equations for a four (or six) dimensional space, where each "point" occurs at the intersection of a pair of two-space (or three-space) points, and the output of the function on that point in the four-space or six-space gives the weight of a connection between them. Inputs and Outputs are at predefined points in the two-space or three-space.
These equations can be solved at various resolutions, allowing the creation of networks to proceed from coarser to finer resolution instead of attempting to train an enormous network all at once.
There are variants of this scheme, of course, mostly in the form of restrictions on the equations.