The output of Deep Neural Networks (DNN) can be altered by a small perturbation of the input in a black box setting by making multiple calls to the DNN. However, the high computation and time required makes the existing approaches unusable. This work seeks to improve the One-pixel (few-pixel) black-box adversarial attacks to reduce the number of calls to the network under attack. The One-pixel attack uses a non-gradient optimization algorithm to find pixel-level perturbations under the constraint of a fixed number of pixels, which causes the network to predict the wrong label for a given image. We show through experimental results how the choice of the optimization algorithm and initial positions to search can reduce function calls and increase attack success significantly, making the attack more practical in real-world settings.

翻译：深神经网络(DNN) 的输出可以通过对 DNN 进行多次调用,对黑盒设置中的输入进行小扰动,对 DNN 进行多次调用,从而改变 DNN 的输出。 但是, 高计算和所需时间使得现有方法无法使用。 这项工作旨在改进一个像素( few- 像素) 黑盒对称攻击, 以减少对网络的呼叫次数。 一像素攻击使用非渐进式优化算法, 在固定数像素的制约下找到像素级的振动, 这使得网络预测给定图像的错误标签。 我们通过实验结果显示, 优化算法和初始位置的搜索选择可以减少功能调用并大大增加攻击成功率, 使攻击在现实世界环境中更加实用。