اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدBio-Inspired Adversarial Attack Against Deep Neural Networks
167
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
The paper develops a new adversarial attack against deep neural networks (DNN), based on applying bio-inspired design to moving physical objects. To the best of our knowledge, this is the first work to introduce physical attacks with a moving object. Instead of following the dominating attack strategy in the existing literature, i.e., to introduce minor perturbations to a digital input or a stationary physical object, we show two new successful attack strategies in this paper. We show by superimposing several patterns onto one physical object, a DNN becomes confused and picks one of the patterns to assign a class label. Our experiment with three flapping wing robots demonstrates the possibility of developing an adversarial camouflage to cause a targeted mistake by DNN. We also show certain motion can reduce the dependency among consecutive frames in a video and make an object detector blind, i.e., not able to detect an object exists in the video. Hence in a successful physical attack against DNN, targeted motion against the system should also be considered.
قيم البحث
اقرأ أيضاً
Although deep neural networks (DNNs) have achieved a great success in various computer vision tasks, it is recently found that they are vulnerable to adversarial attacks. In this paper, we focus on the so-called textit{backdoor attack}, which injects
a backdoor trigger to a small portion of training data (also known as data poisoning) such that the trained DNN induces misclassification while facing examples with this trigger. To be specific, we carefully study the effect of both real and synthetic backdoor attacks on the internal response of vanilla and backdoored DNNs through the lens of Gard-CAM. Moreover, we show that the backdoor attack induces a significant bias in neuron activation in terms of the $ell_infty$ norm of an activation map compared to its $ell_1$ and $ell_2$ norm. Spurred by our results, we propose the textit{$ell_infty$-based neuron pruning} to remove the backdoor from the backdoored DNN. Experiments show that our method could effectively decrease the attack success rate, and also hold a high classification accuracy for clean images.
Machine learning-based malware detection is known to be vulnerable to adversarial evasion attacks. The state-of-the-art is that there are no effective defenses against these attacks. As a response to the adversarial malware classification challenge o
rganized by the MIT Lincoln Lab and associated with the AAAI-19 Workshop on Artificial Intelligence for Cyber Security (AICS2019), we propose six guiding principles to enhance the robustness of deep neural networks. Some of these principles have been scattered in the literature, but the others are introduced in this paper for the first time. Under the guidance of these six principles, we propose a defense framework to enhance the robustness of deep neural networks against adversarial malware evasion attacks. By conducting experiments with the Drebin Android malware dataset, we show that the framework can achieve a 98.49% accuracy (on average) against grey-box attacks, where the attacker knows some information about the defense and the defender knows some information about the attack, and an 89.14% accuracy (on average) against the more capable white-box attacks, where the attacker knows everything about the defense and the defender knows some information about the attack. The framework wins the AICS2019 challenge by achieving a 76.02% accuracy, where neither the attacker (i.e., the challenge organizer) knows the framework or defense nor we (the defender) know the attacks. This gap highlights the importance of knowing about the attack.
In this work, we show how to jointly exploit adversarial perturbation and model poisoning vulnerabilities to practically launch a new stealthy attack, dubbed AdvTrojan. AdvTrojan is stealthy because it can be activated only when: 1) a carefully craft
ed adversarial perturbation is injected into the input examples during inference, and 2) a Trojan backdoor is implanted during the training process of the model. We leverage adversarial noise in the input space to move Trojan-infected examples across the model decision boundary, making it difficult to detect. The stealthiness behavior of AdvTrojan fools the users into accidentally trust the infected model as a robust classifier against adversarial examples. AdvTrojan can be implemented by only poisoning the training data similar to conventional Trojan backdoor attacks. Our thorough analysis and extensive experiments on several benchmark datasets show that AdvTrojan can bypass existing defenses with a success rate close to 100% in most of our experimental scenarios and can be extended to attack federated learning tasks as well.
We study black-box adversarial attacks for image classifiers in a constrained threat model, where adversaries can only modify a small fraction of pixels in the form of scratches on an image. We show that it is possible for adversaries to generate loc
alized textit{adversarial scratches} that cover less than $5%$ of the pixels in an image and achieve targeted success rates of $98.77%$ and $97.20%$ on ImageNet and CIFAR-10 trained ResNet-50 models, respectively. We demonstrate that our scratches are effective under diverse shapes, such as straight lines or parabolic Baezier curves, with single or multiple colors. In an extreme condition, in which our scratches are a single color, we obtain a targeted attack success rate of $66%$ on CIFAR-10 with an order of magnitude fewer queries than comparable attacks. We successfully launch our attack against Microsofts Cognitive Services Image Captioning API and propose various mitigation strategies.
As deep learning systems are widely adopted in safety- and security-critical applications, such as autonomous vehicles, banking systems, etc., malicious faults and attacks become a tremendous concern, which potentially could lead to catastrophic cons
equences. In this paper, we initiate the first study of leveraging physical fault injection attacks on Deep Neural Networks (DNNs), by using laser injection technique on embedded systems. In particular, our exploratory study targets four widely used activation functions in DNNs development, that are the general main building block of DNNs that creates non-linear behaviors -- ReLu, softmax, sigmoid, and tanh. Our results show that by targeting these functions, it is possible to achieve a misclassification by injecting faults into the hidden layer of the network. Such result can have practical implications for real-world applications, where faults can be introduced by simpler means (such as altering the supply voltage).
الأسئلة المقترحة
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
