CIFS: Improving Adversarial Robustness of CNNs via Channel-wise Importance-based Feature Selection

by   Hanshu Yan, et al.

We investigate the adversarial robustness of CNNs from the perspective of channel-wise activations. By comparing non-robust (normally trained) and robustified (adversarially trained) models, we observe that adversarial training (AT) robustifies CNNs by aligning the channel-wise activations of adversarial data with those of their natural counterparts. However, the channels that are negatively-relevant (NR) to predictions are still over-activated when processing adversarial data. Besides, we also observe that AT does not result in similar robustness for all classes. For the robust classes, channels with larger activation magnitudes are usually more positively-relevant (PR) to predictions, but this alignment does not hold for the non-robust classes. Given these observations, we hypothesize that suppressing NR channels and aligning PR ones with their relevances further enhances the robustness of CNNs under AT. To examine this hypothesis, we introduce a novel mechanism, i.e., Channel-wise Importance-based Feature Selection (CIFS). The CIFS manipulates channels' activations of certain layers by generating non-negative multipliers to these channels based on their relevances to predictions. Extensive experiments on benchmark datasets including CIFAR10 and SVHN clearly verify the hypothesis and CIFS's effectiveness of robustifying CNNs.


page 3

page 12


Exploiting Channel Similarity for Accelerating Deep Convolutional Neural Networks

To address the limitations of existing magnitude-based pruning algorithm...

Improving Adversarial Robustness via Channel-wise Activation Suppressing

The study of adversarial examples and their activation has attracted sig...

Analysis and Applications of Class-wise Robustness in Adversarial Training

Adversarial training is one of the most effective approaches to improve ...

Channel redundancy and overlap in convolutional neural networks with channel-wise NNK graphs

Feature spaces in the deep layers of convolutional neural networks (CNNs...

Improving Robustness of Convolutional Neural Networks Using Element-Wise Activation Scaling

Recent works reveal that re-calibrating the intermediate activation of a...

Robustness May Be at Odds with Fairness: An Empirical Study on Class-wise Accuracy

Recently, convolutional neural networks (CNNs) have made significant adv...

Channel-Wise Early Stopping without a Validation Set via NNK Polytope Interpolation

State-of-the-art neural network architectures continue to scale in size ...