Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

What does network.activate do with liquid state machines? #85

Open
sitmo opened this issue Apr 17, 2016 · 1 comment
Open

What does network.activate do with liquid state machines? #85

sitmo opened this issue Apr 17, 2016 · 1 comment

Comments

@sitmo
Copy link

sitmo commented Apr 17, 2016 •
edited
Loading

I was wondering what network.activate does with a random liquid state machines topology.

Suppose I have 1 input and 1 output, and both both are connected to the 2 neurons (A,B) in the hidden layer. Additionally A and B are connected to each other.

layout

Will network.activate()..

  • feed I to A and B
  • also feed the output of (the previous activate() call of) A to B and visa versa
  • compute the output of A and B and then feed it to O?

If so, where does is keep the output state of the hidden neurons A,B? Can you access them? And will .standalone() keep track of those hidden states between calls?
@cazala
Copy link
Owner

cazala commented May 4, 2016

Sure, they will get the last activation of their input neurons. standalone should work fine. When the network is optimized (which is the behaviour by default) everything get stored in a huge array. If you want to access his data after training, you have to call myNetwork.restore() to pull all the data from the optimized array to the actual neuron instances. After this you can just navigate to the desired layer and neuron, and there you have values like activation which hold the last activation of that neuron.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@sitmo @cazala