Collected sources and patterns will appear here. Add from search, explore, or the patterns library.
Tensor<Activation> -> Tensor<DifferentiableGradient>
Approximate the non-differentiable step function of a spiking neuron with a smooth surrogate gradient during backpropagation.
Problem it solves
Spiking activations are step functions whose zero-gradient everywhere blocks backpropagation.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.