Adversarial Co-Evolution of Malware and Detection Models: A Bilevel Optimization Perspective

TL;DR

This paper introduces a bilevel optimization framework for malware detection that models the strategic co-evolution of attackers and defenders, significantly reducing evasion rates and increasing attacker query costs.

Contribution

It presents a novel bilevel optimization approach to enhance malware detection robustness against adaptive adversaries using reinforcement learning.

Findings

Standard classifiers have evasion rates up to 90%.

The proposed method reduces evasion rates to below 2%.

It increases attacker query complexity by up to two orders of magnitude.

Abstract

Machine learning-based malware detectors are increasingly vulnerable to adversarial examples. Traditional defenses, such as one-shot adversarial training, often fail against adaptive attackers who use reinforcement learning to bypass detection. This paper proposes a robust defense framework based on bilevel optimization, explicitly modeling the strategic interaction between a defender and an attacker as an adversarial co-evolutionary process. We evaluate our approach using the MAB-malware framework against three distinct malware families: Mokes, Strab, and DCRat. Our experimental results demonstrate that while standard classifiers and basic adversarial retraining often remain vulnerable, showing evasion rates as high as 90 %, the proposed bilevel optimization approach consistently achieves near-total immunity, reducing evasion rates to 0 - 1.89 %. Furthermore, the iterative framework significantly increases the attacker's query complexity, raising the average cost of successful evasion by up to two orders of magnitude. These findings suggest that modeling the iterative cycle of attack and defense through bilevel optimization is essential for developing resilient malware detection systems capable of withstanding evolving adversarial threats.