DeepAI AI Chat
Log In Sign Up

RANDOM MASK: Towards Robust Convolutional Neural Networks

by   Tiange Luo, et al.
University of Southern California
Peking University

Robustness of neural networks has recently been highlighted by the adversarial examples, i.e., inputs added with well-designed perturbations which are imperceptible to humans but can cause the network to give incorrect outputs. In this paper, we design a new CNN architecture that by itself has good robustness. We introduce a simple but powerful technique, Random Mask, to modify existing CNN structures. We show that CNN with Random Mask achieves state-of-the-art performance against black-box adversarial attacks without applying any adversarial training. We next investigate the adversarial examples which 'fool' a CNN with Random Mask. Surprisingly, we find that these adversarial examples often 'fool' humans as well. This raises fundamental questions on how to define adversarial examples and robustness properly.


page 4

page 12

page 17


Defective Convolutional Layers Learn Robust CNNs

Robustness of convolutional neural networks has recently been highlighte...

Generalized Adversarial Examples: Attacks and Defenses

Most of the works follow such definition of adversarial example that is ...

Retrieval-Augmented Convolutional Neural Networks for Improved Robustness against Adversarial Examples

We propose a retrieval-augmented convolutional network and propose to tr...

Copycat CNN: Stealing Knowledge by Persuading Confession with Random Non-Labeled Data

In the past few years, Convolutional Neural Networks (CNNs) have been ac...

Does Saliency-Based Training bring Robustness for Deep Neural Networks in Image Classification?

Deep Neural Networks are powerful tools to understand complex patterns a...

Defensive Collaborative Multi-task Training - Defending against Adversarial Attack towards Deep Neural Networks

Deep neural network (DNNs) has shown impressive performance on hard perc...

Towards Noise-Robust Neural Networks via Progressive Adversarial Training

Adversarial examples, intentionally designed inputs tending to mislead d...