Convolutional Neural Networks With Dynamic Regularization

Regularization is commonly used for alleviating overfitting in machine learning. For convolutional neural networks (CNNs), regularization methods, such as DropBlock and Shake-Shake, have illustrated the improvement generalization performance. However, these methods lack a self-adaptive ability throughout training. That is, strength fixed to predefined schedule, manual adjustments are required adapt various network architectures. In this article, we propose dynamic method CNNs. Specifically, model function of training loss. According change loss, our can dynamically adjust procedure, thereby balancing underfitting With regularization, large-scale automatically regularized by strong perturbation, vice versa. Experimental results show that proposed improve capability on off-the-shelf architectures outperform state-of-the-art methods.