Key Laboratory of the Ministry of Education for Image Processing and Intelligent Control, School of Artificial Intelligence and Automation, Huazhong University of Science and Technology, Wuhan 430074, People's Republic of China.
School of Design and Environment, National University of Singapore, 117566 Singapore.
J Neural Eng. 2021 Jul 15;18(4). doi: 10.1088/1741-2552/ac0f4c.
. Multiple convolutional neural network (CNN) classifiers have been proposed for electroencephalogram (EEG) based brain-computer interfaces (BCIs). However, CNN models have been found vulnerable to universal adversarial perturbations (UAPs), which are small and example-independent, yet powerful enough to degrade the performance of a CNN model, when added to a benign example.. This paper proposes a novel total loss minimization (TLM) approach to generate UAPs for EEG-based BCIs.. Experimental results demonstrated the effectiveness of TLM on three popular CNN classifiers for both target and non-target attacks. We also verified the transferability of UAPs in EEG-based BCI systems.. To our knowledge, this is the first study on UAPs of CNN classifiers in EEG-based BCIs. UAPs are easy to construct, and can attack BCIs in real-time, exposing a potentially critical security concern of BCIs.
. 已经提出了多种卷积神经网络（CNN）分类器用于基于脑电图（EEG）的脑机接口（BCI）。然而，CNN 模型被发现容易受到通用对抗扰动（UAPs）的影响，UAPs 是小的、与示例无关的，但足以降低 CNN 模型的性能，当添加到良性示例时。本文提出了一种新的总损失最小化（TLM）方法来生成基于 EEG 的 BCI 中的 UAPs。实验结果表明，TLM 对三种流行的 CNN 分类器在目标和非目标攻击中都有效。我们还验证了 UAPs 在基于 EEG 的 BCI 系统中的可转移性。据我们所知，这是首次对基于 EEG 的 BCI 中 CNN 分类器的 UAPs 进行研究。UAPs 易于构建，并且可以实时攻击 BCI，这暴露出 BCI 潜在的关键安全问题。
