Periodic Activation Functions Induce Stationarity

Lassi Meronen · Martin Trapp · Arno Solin

Keywords: [ Deep Learning ] [ Kernel Methods ]

[ Abstract ]
Thu 9 Dec 8:30 a.m. PST — 10 a.m. PST


Neural network models are known to reinforce hidden data biases, making them unreliable and difficult to interpret. We seek to build models that `know what they do not know' by introducing inductive biases in the function space. We show that periodic activation functions in Bayesian neural networks establish a connection between the prior on the network weights and translation-invariant, stationary Gaussian process priors. Furthermore, we show that this link goes beyond sinusoidal (Fourier) activations by also covering triangular wave and periodic ReLU activation functions. In a series of experiments, we show that periodic activation functions obtain comparable performance for in-domain data and capture sensitivity to perturbed inputs in deep neural networks for out-of-domain detection.

Chat is not available.