How to Train your Antivirus: RL-based Hardening through the Problem-Space

Abstract

ML-based malware detection on dynamic analysis reports is vulnerable to both evasion and spurious correlations. In this work, we investigate a specific ML architecture employed in the pipeline of a widely-known commercial antivirus company, with the goal to harden it against adversarial malware. Adversarial training, the sole defensive technique that can confer empirical robustness, is not applicable out of the box in this domain, for the principal reason that gradient-based perturbations rarely map back to feasible problem-space programs. We introduce a novel Reinforcement Learning approach for constructing adversarial examples, a constituent part of adversarially training a model against evasion. Our approach comes with multiple advantages. It performs modifications that are feasible in the problem-space, and only those; thus it circumvents the inverse mapping problem. It also makes possible to provide theoretical guarantees on the robustness of the model against a particular set of adversarial capabilities. Our empirical exploration validates our theoretical insights, where we can consistently reach 0% Attack Success Rate after a few adversarial retraining iterations.

20 pages,4 figures

References in corpus (9)

• Shortcut Learning in Deep Neural Networks
• Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)
• A General Framework for Adversarial Examples with Objectives
• Functionality-preserving Black-box Optimization of Adversarial Windows Malware
• Adversarial EXEmples: A Survey and Experimental Evaluation of Practical Attacks on Machine Learning for Windows Malware Detection
• A Framework for Enhancing Deep Neural Networks Against Adversarial Malware
• Towards Robust and Reliable Algorithmic Recourse
• A survey on practical adversarial examples for malware classifiers
• Robust Feature-Level Adversaries are Interpretability Tools