Abstract

Over the past few years, softmax and SGD have become a commonly used component and the default training strategy in CNN frameworks, respectively. However, when optimizing CNNs with SGD, the saturation behavior behind softmax always gives us an illusion of training well and then is omitted. In this paper, we fifirst emphasize that the early saturation behavior of softmax will impede the exploration of SGD, which sometimes is a reason for model converging at a bad local-minima, then propose Noisy Softmax to mitigating this early saturation issue by injecting annealed noise in softmax during each iteration. This operation based on noise injection aims at postponing the early saturation and further bringing continuous gradients propagation so as to signifificantly encourage SGD solver to be more exploratory and help to fifind a better local-minima. This paper empirically verififies the superiority of the early softmax desaturation, and our method indeed improves the generalization ability of CNN model by regularization. We experimentally fifind that this early desaturation helps optimization in many tasks, yielding state-of-the-art or competitive results on several popular benchmark datasets.