Open Seanny123 opened 8 years ago
I see no reason why not! In the future we'll be making some infrastructure to let people put networks, examples, etc in other repositories and making them discoverable (by the GUI at least), and at that point we might make a repo specifically for networks like this, but there's no reason to wait for that point to make the network available here.
While I'm at it, I should also add a point attractor.
I've also got a collection of random attractors here
https://github.com/tcstewar/testing_notebooks/blob/master/Attractors%20Collection.ipynb
(note: those examples were done in the rather extreme domain of having sigmoid neurons and getting rid of the noise on the decoder solver, to try to see how few neurons I could get away with....)
@tcstewar which of those attractors have been previously used in a Nengo application? I ask, because the reason I'm making reference implementations of point attractors and frequency controlled oscillators is their importance for the wildly extensible Dynamic Movement Primitives, which are used for motor control. I figure the dynamical systems that are most commonly used is where I should focus my efforts.
@tcstewar which of those attractors have been previously used in a Nengo application?
None of them. They're basically just gimmicky examples, along the lines of the Lorenz attractor (which we've also never really used in a really application, other than showing off)... well, technically, I guess I did use the Lorenz attractor to control a robot for a demo, but still not a real example. :)
This is used a lot in control systems. Would adding this as a network to this repository be acceptable? I'm assuming it shouldn't go into core nengo.