Smooth Inter-layer Propagation of Stabilized Neural Networks for Classification
This work addresses the theoretical understanding of ResNet performance for researchers in deep learning, but it appears incremental as it builds on existing studies without introducing new methods or data.
The paper investigates the role of stability and smoothness in inter-layer propagation of ResNets, aiming to explain when these properties enhance performance in image classification, with a focus on batch normalization and dynamical systems views.
Recent work has studied the reasons for the remarkable performance of deep neural networks in image classification. We examine batch normalization on the one hand and the dynamical systems view of residual networks on the other hand. Our goal is in understanding the notions of stability and smoothness of the inter-layer propagation of ResNets so as to explain when they contribute to significantly enhanced performance. We postulate that such stability is of importance for the trained ResNet to transfer.