An Adversarial Approach for Intrusion Detection Systems Using Jacobian Saliency Map Attacks (JSMA) Algorithm

Qureshi, Ayyaz-Ul-Haq; Larijani, Hadi; Yousefi, Mohammadmehdi; Adeel, Ahsan; Mtetwa, Nhamoinesu

doi:10.3390/computers9030058

Cited by 18 publications

(5 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Random Neural Network-ADV [52] method the authors have used adversarial retraining approach to improve the robustness of random neural network. NSL KDD dataset is used for the experiments.…”

Section: A Proactivementioning

confidence: 99%

Defenses for Adversarial attacks in Network Intrusion Detection System – A Survey

Dhinakaran¹,

Anto²

2023

IJCDS

View full text Add to dashboard Cite

In computer security, machine learning has a greater impact in recent years. Ranging from spam filtering, malware analysis, and traffic analysis to network security the usage of machine learning algorithms are manifold. In the area of network security, machine learning techniques are used especially in developing intrusion detection systems. There are basically two kinds of intrusion detection systems -host intrusion detection systems and network intrusion detection systems. Even though machine learning techniques have greatly improved the efficiency of the intrusion detection systems, they are vulnerable to adversarial attacks which are designed and launched by adaptive adversaries who know the working principles of machine learning models. In recent years adversarial machine learning has gained attention in the domain of machine learning in which attackers exploit the inherent fallacies in the assumptions made in the machine learning models. In the domain of network security especially in intrusion detection systems, the significant role of adversarial machine learning has not been addressed in detail. This survey examines different types of defenses deployed to mitigate the impact of adversarial attacks. Their effectiveness in dealing with attacks is analysed and their limitations are discussed.

show abstract

Section: A Proactivementioning

confidence: 99%

Defenses for Adversarial attacks in Network Intrusion Detection System – A Survey

Dhinakaran¹,

Anto²

2023

IJCDS

View full text Add to dashboard Cite

show abstract

“…Qureshi et al proposed a novel adversarial intrusion detection system based on random neural networks (RNN-ADV). However, the perturbation environment significantly affects the performance of this model, which performs better in terms of accuracy and F1-score compared to deep neural networks when using the JSMA algorithm 30 . In the literature 31 , Safa et al proposed a joint reinforcement learning-based intrusion detection system (FRL-IDS) for Internet of Things (IoT) networks in healthcare infrastructure.…”

Section: Related Workmentioning

confidence: 99%

ID-RDRL: a deep reinforcement learning-based feature selection intrusion detection model

Ren

Zeng

Cao

et al. 2022

Sci Rep

View full text Add to dashboard Cite

Network assaults pose significant security concerns to network services; hence, new technical solutions must be used to enhance the efficacy of intrusion detection systems. Existing approaches pay insufficient attention to data preparation and inadequately identify unknown network threats. This paper presents a network intrusion detection model (ID-RDRL) based on RFE feature extraction and deep reinforcement learning. ID-RDRL filters the optimum subset of features using the RFE feature selection technique, feeds them into a neural network to extract feature information and then trains a classifier using DRL to recognize network intrusions. We utilized CSE-CIC-IDS2018 as a dataset and conducted tests to evaluate the model’s performance, which is comprised of a comprehensive collection of actual network traffic. The experimental results demonstrate that the proposed ID-RDRL model can select the optimal subset of features, remove approximately 80% of redundant features, and learn the selected features through DRL to enhance the IDS performance for network attack identification. In a complicated network environment, it has promising application potential in IDS.

show abstract

“…In whitebox attacks, the attacker has knowledge of the attacked model's internal structure and parameters, and exploits that knowledge. It is commonly performed by using the model's gradients [5,7,12].…”

Section: Related Workmentioning

confidence: 99%

An Evolutionary, Gradient-Free, Query-Efficient, Black-Box Algorithm for Generating Adversarial Instances in Deep Convolutional Neural Networks

2022

View full text Add to dashboard Cite

Deep neural networks (DNNs) are sensitive to adversarial data in a variety of scenarios, including the black-box scenario, where the attacker is only allowed to query the trained model and receive an output. Existing black-box methods for creating adversarial instances are costly, often using gradient estimation or training a replacement network. This paper introduces Qu ery-Efficient Evolutionary Attack—QuEry Attack—an untargeted, score-based, black-box attack. QuEry Attack is based on a novel objective function that can be used in gradient-free optimization problems. The attack only requires access to the output logits of the classifier and is thus not affected by gradient masking. No additional information is needed, rendering our method more suitable to real-life situations. We test its performance with three different, commonly used, pretrained image-classifications models—Inception-v3, ResNet-50, and VGG-16-BN—against three benchmark datasets: MNIST, CIFAR10 and ImageNet. Furthermore, we evaluate QuEry Attack’s performance on non-differential transformation defenses and robust models. Our results demonstrate the superior performance of QuEry Attack, both in terms of accuracy score and query efficiency.

show abstract

An Adversarial Approach for Intrusion Detection Systems Using Jacobian Saliency Map Attacks (JSMA) Algorithm

Cited by 18 publications

References 19 publications

Defenses for Adversarial attacks in Network Intrusion Detection System – A Survey

Defenses for Adversarial attacks in Network Intrusion Detection System – A Survey

ID-RDRL: a deep reinforcement learning-based feature selection intrusion detection model

An Evolutionary, Gradient-Free, Query-Efficient, Black-Box Algorithm for Generating Adversarial Instances in Deep Convolutional Neural Networks

Contact Info

Product

Resources

About