-
Notifications
You must be signed in to change notification settings - Fork 0
Generalize recurrent-to-readout and readout-to-recurrent delays #31
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.
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
base: eprop_bio_feature
Are you sure you want to change the base?
Conversation
pynest/examples/eprop_plasticity/eprop_supervised_regression_sine-waves.py
Outdated
Show resolved
Hide resolved
pynest/examples/eprop_plasticity/eprop_supervised_regression_sine-waves.py
Outdated
Show resolved
Hide resolved
pynest/examples/eprop_plasticity/eprop_supervised_regression_sine-waves.py
Outdated
Show resolved
Hide resolved
pynest/examples/eprop_plasticity/eprop_supervised_regression_sine-waves.py
Outdated
Show resolved
Hide resolved
pynest/examples/eprop_plasticity/eprop_supervised_regression_sine-waves.py
Outdated
Show resolved
Hide resolved
pynest/examples/eprop_plasticity/eprop_supervised_regression_sine-waves.py
Outdated
Show resolved
Hide resolved
pynest/examples/eprop_plasticity/eprop_supervised_regression_sine-waves.py
Outdated
Show resolved
Hide resolved
pynest/examples/eprop_plasticity/eprop_supervised_classification_neuromorphic_mnist.py
Outdated
Show resolved
Hide resolved
pynest/examples/eprop_plasticity/eprop_supervised_classification_neuromorphic_mnist.py
Outdated
Show resolved
Hide resolved
pynest/examples/eprop_plasticity/eprop_supervised_classification_neuromorphic_mnist.py
Outdated
Show resolved
Hide resolved
pynest/examples/eprop_plasticity/eprop_supervised_classification_neuromorphic_mnist.py
Outdated
Show resolved
Hide resolved
pynest/examples/eprop_plasticity/eprop_supervised_classification_neuromorphic_mnist.py
Outdated
Show resolved
Hide resolved
|
973e0bb to
b1d0289
Compare
b1d0289 to
0197878
Compare
|
Pull request automatically marked stale! |
Enhance e-prop plasticity with biologically inspired features
…o-recurrent layers.
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
* add support for generalized delays to the evidence accumulation task
0197878 to
aa13386
Compare
akorgor
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you for this nice feature! As discussed this would be ultimately best as a NEST extension module.
| if ( delay_out_rec_ < 1 ) | ||
| { | ||
| throw BadProperty( "Broadcast delay of learning signals ≥ 1 required." ); | ||
| } | ||
|
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
| if ( delay_out_rec_ < 1 ) | |
| { | |
| throw BadProperty( "Broadcast delay of learning signals ≥ 1 required." ); | |
| } | |
| if ( delay_out_rec_ < 1 ) | |
| { | |
| throw BadProperty( "Connection delay from readout to recurrent neuron ≥ 1 required." ); | |
| } | |
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]> Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
Co-authored-by: Agnes Korcsak-Gorzo <[email protected]>
|
Pull request automatically marked stale! |
This PR introduces modifications to handle the propagation delay of spikes propagating from the recurrent layer to the readout layer when this delay surpasses one resolution step. It also supports learning signals with delays exceeding one resolution step.
Additionally, the PR implements a series of consistency checks to ensure the proper configuration of generalized delay parameters (
d_rec_outandd_out_rec) across various types of eprop neurons throughout the neural network.Consistency Check for Recurrent and Output Neurons (
d_out_rec):eprop_iaf,eprop_iaf_adapt,eprop_iaf_psc_delta) transmit a test event using a delay specified by their ownd_rec_outvalue.eprop_readout) receive the test event and compare the actual delay against their ownd_rec_outparameter.Connection-Specific Consistency Checks:
Rec-Rec Connection:
d_rec_outandd_out_rec) from the target recurrent neuron.Rec-Out Connection:
delayis verified to match thed_rec_outset by the target output neuron using thecheck_connectionfunction.Out-Rec Connection (Feedback Connection):
delayis verified to match thed_out_recset by the target output neuron using thecheck_connectionfunction.The following diagram illustrates these consistency checks.
