Explaining Black-box Android Malware Detection

TL;DR

This paper introduces a gradient-based method to interpret black-box Android malware detection models, enhancing understanding of their decisions and vulnerabilities, and addressing the fragility of traditional benchmark evaluations.

Contribution

It generalizes explainability techniques to any black-box model, allowing for improved interpretability and vulnerability analysis in Android malware detection.

Findings

Effective identification of influential features for any model

Reveals global model characteristics for malware detection

Highlights potential adversarial vulnerabilities

Abstract

Machine-learning models have been recently used for detecting malicious Android applications, reporting impressive performances on benchmark datasets, even when trained only on features statically extracted from the application, such as system calls and permissions. However, recent findings have highlighted the fragility of such in-vitro evaluations with benchmark datasets, showing that very few changes to the content of Android malware may suffice to evade detection. How can we thus trust that a malware detector performing well on benchmark data will continue to do so when deployed in an operating environment? To mitigate this issue, the most popular Android malware detectors use linear, explainable machine-learning models to easily identify the most influential features contributing to each decision. In this work, we generalize this approach to any black-box machine- learning model, by leveraging a gradient-based approach to identify the most influential local features. This enables using nonlinear models to potentially increase accuracy without sacrificing interpretability of decisions. Our approach also highlights the global characteristics learned by the model to discriminate between benign and malware applications. Finally, as shown by our empirical analysis on a popular Android malware detection task, it also helps identifying potential vulnerabilities of linear and nonlinear models against adversarial manipulations.