Statically Detecting Adversarial Malware through Randomised Chaining

TL;DR

This paper introduces a randomized chaining approach to detect adversarial malware attacks, aiming to improve static detection robustness against evasion techniques in machine learning-based antivirus systems.

Contribution

It proposes a novel randomized chaining method to enhance static malware detection and defend against adversarial evasion attacks.

Findings

The method effectively detects adversarial malware in static analysis.

Randomized chaining increases robustness against evasion techniques.

The approach contributes to more secure malware detection systems.

Abstract

With the rapid growth of malware attacks, more antivirus developers consider deploying machine learning technologies into their productions. Researchers and developers published various machine learning-based detectors with high precision on malware detection in recent years. Although numerous machine learning-based malware detectors are available, they face various machine learning-targeted attacks, including evasion and adversarial attacks. This project explores how and why adversarial examples evade malware detectors, then proposes a randomised chaining method to defend against adversarial malware statically. This research is crucial for working towards combating the pertinent malware cybercrime.