Denial-of-Service Attack on Object Detection Model Using Universal Adversarial Perturbation

Shapira, Avishag; Zolfi, Alon; Demetrio, Luca; Biggio, Battista; Shabtai, Asaf

doi:10.48550/arxiv.2205.13618

Cited by 2 publications

(2 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…End to End time This one path direction calculated as time from source to destination. [99] Packet drop…”

Section: Objective Functions Description Referencesmentioning

confidence: 99%

High-Speed Network DDoS Attack Detection: A Survey

Haseeb-ur-rehman,

Aman,

Hasan

et al. 2023

Sensors

View full text Add to dashboard Cite

Having a large number of device connections provides attackers with multiple ways to attack a network. This situation can lead to distributed denial-of-service (DDoS) attacks, which can cause fiscal harm and corrupt data. Thus, irregularity detection in traffic data is crucial in detecting malicious behavior in a network, which is essential for network security and the integrity of modern Cyber–Physical Systems (CPS). Nevertheless, studies have shown that current techniques are ineffective at detecting DDoS attacks on networks, especially in the case of high-speed networks (HSN), as detecting attacks on the latter is very complex due to their fast packet processing. This review aims to study and compare different approaches to detecting DDoS attacks, using machine learning (ML) techniques such as k-means, K-Nearest Neighbors (KNN), and Naive Bayes (NB) used in intrusion detection systems (IDSs) and flow-based IDSs, and expresses data paths for packet filtering for HSN performance. This review highlights the high-speed network accuracy evaluation factors, provides a detailed DDoS attack taxonomy, and classifies detection techniques. Moreover, the existing literature is inspected through a qualitative analysis, with respect to the factors extracted from the presented taxonomy of irregular traffic pattern detection. Different research directions are suggested to support researchers in identifying and designing the optimal solution by highlighting the issues and challenges of DDoS attacks on high-speed networks.

show abstract

“…End to End time This one path direction calculated as time from source to destination. [99] Packet drop…”

Section: Objective Functions Description Referencesmentioning

confidence: 99%

High-Speed Network DDoS Attack Detection: A Survey

Haseeb-ur-rehman,

Aman,

Hasan

et al. 2023

Sensors

View full text Add to dashboard Cite

show abstract

“…Weng et al [47] leverage the Kullback-Leibler (KL) divergence loss to implement both targeted and non-targeted universal attacks. And universal adversarial attacks can be adopted to different tasks such as remote sensing [48], text recognition [49], watermarking [50], object detection [51] and so on.…”

Section: Image-dependent Attacksmentioning

confidence: 99%

Crafting Targeted Universal Adversarial Perturbations: Considering Images as Noise

Wang,

Cai,

Wang

et al. 2023

IEEE Access

View full text Add to dashboard Cite

The vulnerability of deep neural networks to adversarial perturbations has been demonstrated in a large body of research. Compared to image-dependent adversarial perturbations, universal adversarial perturbations(UAPs) is more challenging for indiscriminately attacking the model inputs. However, there are few studies on generating data-free targeted UAPs and the targeted attack success rate of the latest method remains unsatisfactory. Not only that, fewer studies have implemented their approach on Transformers and its efficacy remains uncertain. Therefore, a novel method denoted as Denoising Targeted UAP (DT-UAP) is proposed in this paper that considers the training input as the noise, and takes the input of the last layer into calculation. Specifically, the proposed method minimizes the distance between perturbations and adversarial examples, then incorporates a targeted loss function to generate targeted universal adversarial perturbations for different DNNs and Transformers based on different proxy datasets. DT-UAP has achieved an average improvement of 5% to 10% in terms of both fooling rate and targeted fooling rate comparing to the most recent method for generating targeted universal adversarial perturbation with proxy dataset for DNNs. Additionally, DT-UAP has also achieved a targeted attack success rate of over 80% on Transformers such as MaxVit and SwinTransformer.INDEX TERMS targeted universal adversarial perturbation, adversarial example, Deep Neural Network, Transformer, image as noise, proxy dataset.

show abstract

Denial-of-Service Attack on Object Detection Model Using Universal Adversarial Perturbation

Cited by 2 publications

References 29 publications

High-Speed Network DDoS Attack Detection: A Survey

High-Speed Network DDoS Attack Detection: A Survey

Crafting Targeted Universal Adversarial Perturbations: Considering Images as Noise

Contact Info

Product

Resources

About