Spectral Norm Regularization for Improving the Generalizability of Deep Learning скачать в хорошем качестве

Spectral Norm Regularization for Improving the Generalizability of Deep Learning

In this video we discuss the lowering the spectral norm of the weight matrix as a regularization technique for neural networks. Paper: https://arxiv.org/abs/1705.10941 Abstract: We investigate the generalizability of deep learning based on the sensitivity to input perturbation. We hypothesize that the high sensitivity to the perturbation of data degrades the performance on it. To reduce the sensitivity to perturbation, we propose a simple and effective regularization method, referred to as spectral norm regularization, which penalizes the high spectral norm of weight matrices in neural networks. We provide supportive evidence for the abovementioned hypothesis by experimentally confirming that the models trained using spectral norm regularization exhibit better generalizability than other baseline methods.