Sensitivity Analysis of Deep Neural Networks

Abstract

Deep neural networks (DNNs) have achieved superior performance in various prediction tasks, but can be very vulnerable to adversarial examples or perturbations. Therefore, it is crucial to measure the sensitivity of DNNs to various forms of perturbations in real applications. We introduce a novel perturbation manifold and its associated influence measure to quantify the effects of various perturbations on DNN classifiers. Such perturbations include various external and internal perturbations to input samples and network parameters. The proposed measure is motivated by information geometry and provides desirable invariance properties. We demonstrate that our influence measure is useful for four model building tasks: detecting potential 'outliers', analyzing the sensitivity of model architectures, comparing network sensitivity between training and test sets, and locating vulnerable areas. Experiments show reasonably good performance of the proposed measure for the popular DNN models ResNet50 and DenseNet121 on CIFAR10 and MNIST datasets.

arXiv commentAccepted by AAAI-19

References in corpus (7)

• TensorFlow: A system for large-scale machine learning
• Intriguing properties of neural networks
• End to End Learning for Self-Driving Cars
• Sensitivity and Generalization in Neural Networks: an Empirical Study
• Formal Guarantees on the Robustness of a Classifier against Adversarial Manipulation
• Evaluating the Robustness of Neural Networks: An Extreme Value Theory Approach
• On the Robustness of Convolutional Neural Networks to Internal Architecture and Weight Perturbations

Cited by in corpus (2)

• mFI-PSO: A Flexible and Effective Method in Adversarial Image Generation for Deep Neural Networks
• The Impact of Hole Geometry on Relative Robustness of In-Painting Networks: An Empirical Study