Empirical study of the modulus as activation function in computer vision applications
This addresses the need for better activation functions in computer vision, particularly for TinyML and hardware applications, though it appears incremental as it builds on existing research on nonlinearities.
The authors tackled the problem of improving generalization in computer vision models by proposing the modulus as a new non-monotonic activation function, resulting in up to a 15% accuracy increase on CIFAR100 and 4% on CIFAR10 compared to benchmark activations.
In this work we propose a new non-monotonic activation function: the modulus. The majority of the reported research on nonlinearities is focused on monotonic functions. We empirically demonstrate how by using the modulus activation function on computer vision tasks the models generalize better than with other nonlinearities - up to a 15% accuracy increase in CIFAR100 and 4% in CIFAR10, relative to the best of the benchmark activations tested. With the proposed activation function the vanishing gradient and dying neurons problems disappear, because the derivative of the activation function is always 1 or -1. The simplicity of the proposed function and its derivative make this solution specially suitable for TinyML and hardware applications.