Random Directional Attack for Fooling Deep Neural Networks

TL;DR

This paper introduces a novel random directional attack (RDA) method that effectively generates adversarial examples for deep neural networks by exploring attack directions beyond gradients, performing well in black-box scenarios.

Contribution

The paper proposes RDA, a new attack method that searches attack directions randomly with hill climbing, outperforming gradient-based methods especially in black-box settings.

Findings

RDA achieves competitive attack success rates.

RDA performs similarly in black-box and white-box attacks.

RDA outperforms existing gradient-based attack methods.

Abstract

Deep neural networks (DNNs) have been widely used in many fields such as images processing, speech recognition; however, they are vulnerable to adversarial examples, and this is a security issue worthy of attention. Because the training process of DNNs converge the loss by updating the weights along the gradient descent direction, many gradient-based methods attempt to destroy the DNN model by adding perturbations in the gradient direction. Unfortunately, as the model is nonlinear in most cases, the addition of perturbations in the gradient direction does not necessarily increase loss. Thus, we propose a random directed attack (RDA) for generating adversarial examples in this paper. Rather than limiting the gradient direction to generate an attack, RDA searches the attack direction based on hill climbing and uses multiple strategies to avoid local optima that cause attack failure. Compared with state-of-the-art gradient-based methods, the attack performance of RDA is very competitive. Moreover, RDA can attack without any internal knowledge of the model, and its performance under black-box attack is similar to that of the white-box attack in most cases, which is difficult to achieve using existing gradient-based attack methods.