High-Frequency Component Helps Explain the Generalization of Convolutional Neural Networks
Haohan Wang, Xindi Wu, Zeyi Huang, Eric P. Xing
Abstract
We investigate the relationship between the frequency spectrum of image data and the generalization behavior of convolutional neural networks (CNN). We first notice CNN's ability in capturing the high-frequency components of images. These high-frequency components are almost imperceptible to a human. Thus the observation leads to multiple hypotheses that are related to the generalization behaviors of CNN, including a potential explanation for adversarial examples, a discussion of CNN's trade-off between robustness and accuracy, and some evidence in understanding training heuristics.
Topics & Concepts
Convolutional neural networkComputer scienceGeneralizationRobustness (evolution)Artificial intelligenceHeuristicsNoticeComponent (thermodynamics)Pattern recognition (psychology)Machine learningMathematicsPhysicsLawThermodynamicsOperating systemChemistryBiochemistryPolitical scienceMathematical analysisGeneAdversarial Robustness in Machine LearningAnomaly Detection Techniques and ApplicationsAdvanced Neural Network Applications