English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Machine learning (ML) has gained significant adoption in Android malwaredetection to address the escalating threats posed by the rapid proliferation ofmalware attacks. However, recent studies have revealed the inherentvulnerabilities of ML-based detection systems to evasion attacks. While effortshave been made to address this critical issue, many of the existing defensivemethods encounter challenges such as lower effectiveness or reducedgeneralization capabilities. In this paper, we introduce a novel Androidmalware detection method, MalPurifier, which exploits adversarial purificationto eliminate perturbations independently, resulting in attack mitigation in alight and flexible way. Specifically, MalPurifier employs a DenoisingAutoEncoder (DAE)-based purification model to preprocess input samples,removing potential perturbations from them and then leading to correctclassification. To enhance defense effectiveness, we propose a diversifiedadversarial perturbation mechanism that strengthens the purification modelagainst different manipulations from various evasion attacks. We alsoincorporate randomized "protective noises" onto benign samples to preventexcessive purification. Furthermore, we customize a loss function for improvingthe DAE model, combining reconstruction loss and prediction loss, to enhancefeature representation learning, resulting in accurate reconstruction andclassification. Experimental results on two Android malware datasetsdemonstrate that MalPurifier outperforms the state-of-the-art defenses, and itsignificantly strengthens the vulnerable malware detector against 37 evasionattacks, achieving accuracies over 90.91%. Notably, MalPurifier demonstrateseasy scalability to other detectors, offering flexibility and robustness in itsimplementation.

MalPurifier: Enhancing Android Malware Detection with Adversarial Purification against Evasion Attacks