Explainable and Transferable Adversarial Attack for ML-Based Network Intrusion Detectors

Despite being widely used in network intrusion detection systems (NIDSs), machine learning (ML) has proven to be vulnerable to adversarial attacks. White-box and black-box adversarial ML attacks of NIDS have been explored in several studies. However, white-box attacks unrealistically assume that the attackers have full knowledge of the target NIDSs. Meanwhile, existing black-box attacks can not achieve high attack success rate due to the weak adversarial transferability between models (e.g., neural networks and tree models). Additionally, neither of them explains why adversarial examples exist and why they can transfer across models. To address these challenges, this paper introduces ETA, an <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">E</u>xplainable <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">T</u>ransfer-based Black-Box Adversarial <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">A</u>ttack framework. ETA aims to achieve two primary objectives: 1) create transferable adversarial examples applicable to various ML detectors and 2) provide insights into the existence of adversarial examples and their transferability within NIDSs. Specifically, we first provide a general transfer-based adversarial attack method applicable across the entire ML space. Following that, we exploit a unique insight based on cooperative game theory and perturbation interpretations to explain adversarial examples and adversarial transferability. On this basis, we propose an Important-Sensitive Feature Selection (ISFS) method to guide the search for adversarial examples, achieving stronger transferability and ensuring traffic-space constraints. Finally, the experimental results on three NIDSs datasets show that our method performs significantly effectively against several classical and state-of-the-art ML classifiers, outperforming the latest baselines. We conduct three interpretation experiments and two cases to verify our interpretation method's correctness. Meanwhile, we uncover two major misconceptions about applying machine learning to NIDSs systems.