Activation Functions

# Shifted Softplus

Introduced by Schütt et al. in SchNet: A continuous-filter convolutional neural network for modeling quantum interactions

Shifted Softplus is an activation function ${\rm ssp}(x) = \ln( 0.5 e^{x} + 0.5 )$, which SchNet employs as non-linearity throughout the network in order to obtain a smooth potential energy surface. The shifting ensures that ${\rm ssp}(0) = 0$ and improves the convergence of the network. This activation function shows similarity to ELUs, while having infinite order of continuity.

#### Papers

Paper Code Results Date Stars