arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2111.09488cs.CRcs.LG
收藏

使用通用对抗性扰动攻击深度学习AI硬件

Attacking Deep Learning AI Hardware with Universal Adversarial Perturbation

Auburn University(奥本大学);Florida International University(佛罗里达国际大学)

Mehdi Sadi, B. M. S. Bahar Talukder, Kaniz Mishty, Md Tauhidur Rahman

摘要:

通用对抗性扰动是与图像无关且与模型无关的噪声,当与任何图像一起添加时,都会误导经过训练的深度卷积神经网络进行错误的预测。由于这些通用对抗扰动可能严重危害实际深度学习应用的安全性和完整性,因此现有技术使用附加的神经网络来检测输入图像源处这些噪声的存在。在本文中,我们展示了一种攻击策略,当被流氓手段激活时(例如,恶意软件、木马)可以通过在AI硬件加速器阶段增加对抗性噪声来绕过这些现有的对策。我们在FuseSoC环境下,通过Conv2D函数的软件内核和硬件的Verilog RTL模型的联合仿真,演示了对几种深度学习模型的加速器级通用对抗性噪声攻击。

英文摘要:

Universal Adversarial Perturbations are image-agnostic and model-independent noise that when added with any image can mislead the trained Deep Convolutional Neural Networks into the wrong prediction. Since these Universal Adversarial Perturbations can seriously jeopardize the security and integrity of practical Deep Learning applications, existing techniques use additional neural networks to detect the existence of these noises at the input image source. In this paper, we demonstrate an attack strategy that when activated by rogue means (e.g., malware, trojan) can bypass these existing countermeasures by augmenting the adversarial noise at the AI hardware accelerator stage. We demonstrate the accelerator-level universal adversarial noise attack on several deep Learning models using co-simulation of the software kernel of Conv2D function and the Verilog RTL model of the hardware under the FuseSoC environment.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑