You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Yes, it is possible to implement regression tasks, this is mostly down to the choice of Loss function. One thing that is tricky and an active area of research is how to best define a loss which only operates on spike times in order to do fit for example the behaviour of a Neuron model to measurements. I can look into adding an example in the documentation.
One way could be to add a linear output layer, which converts a (sufficiently large) number of spikes to numeric values. This wouldn't be "pure", but it should work with standard losses. But I am curious to see other approaches. Looking forward to the examples, maybe with hints to publications?
Best
Is there a way to perform regression tasks with Norse? If so can you kindly provide an example notebook or a link from the documentation.
The text was updated successfully, but these errors were encountered: