Adaptative Perturbation Patterns: Realistic Adversarial Learning for Robust Intrusion Detection

Abstract

Adversarial attacks pose a major threat to machine learning and to the systems that rely on it. In the cybersecurity domain, adversarial cyber-attack examples capable of evading detection are especially concerning. Nonetheless, an example generated for a domain with tabular data must be realistic within that domain. This work establishes the fundamental constraint levels required to achieve realism and introduces the adaptative perturbation pattern method (A2PM) to fulfill these constraints in a gray-box setting. A2PM relies on pattern sequences that are independently adapted to the characteristics of each class to create valid and coherent data perturbations. The proposed method was evaluated in a cybersecurity case study with two scenarios: Enterprise and Internet of Things (IoT) networks. Multilayer perceptron (MLP) and random forest (RF) classifiers were created with regular and adversarial training, using the CIC-IDS2017 and IoT-23 datasets. In each scenario, targeted and untargeted attacks were performed against the classifiers, and the generated examples were compared with the original network traffic flows to assess their realism. The obtained results demonstrate that A2PM provides a scalable generation of realistic adversarial examples, which can be advantageous for both adversarial training and attacks.

Ransomware Detection and Classification Using Generative Adversarial Networks with Dynamic Weight Adaptation

“…GANs excelled in generating diverse data samples, enabling the detection models to learn from a broader spectrum of ransomware behaviors and characteristics, thereby improving detection accuracy [25]. Additionally, GANs proved effective in identifying vulnerabilities within detection systems by generating adversarial examples, which highlighted weaknesses and informed the development of more resilient detection mechanisms [26]. The adaptability of GANs, driven by their ability to learn complex data distributions, allowed for the continuous improvement of detection models as new ransomware variants emerged [27].…”

Adv-Bot: Realistic adversarial botnet attacks against network intrusion detection systems

“…The literature study shows that there has been considerable research on the impact of adversarial attacks on machine learning models [8,9,10]. However, their feasibility in domainconstrained applications, such as intrusion detection systems, is still in its early stages [11,12,13]. Adversarial attacks can be performed in either white-box or black-box settings.…”

Cybersecurity of Critical Infrastructures: Challenges and Solutions

“…A vital challenge faced nowadays by federal and business decision-makers for choosing cost-efficient mitigations to scale back risks from supply chain attacks, particularly those from adversarial attacks that are complex, hard to detect and can lead to severe consequences. Focusing on adversarial attacks and how these can alter the performance of AI based detection systems, the authors in [ 13 ] propose a novel robust solution. Their proposed model was evaluated in both Enterprise and Internet of Things (IoT) networks and is proven to be efficient against adversarial classification attacks and adversarial training attacks.…”