Abstract:Machine learning (ML) in real-world systems must contend with concept drift, adversarial actors, and a spectrum of potential features with varying costs and benefits. Malware naturally exhibits all of these complexities, but for the same reason, it is challenging to curate and organize data to study these factors. We present McNdroid, to our knowledge the largest longitudinal multimodal Android malware benchmark for malware detection and drift analysis. McNdroid spans 2013--2025, excluding 2015, and represents each application with three aligned modalities--static features from manifests and smali code, dynamic behavioral features from sandbox execution, and graph-based features from function-call graphs. Using temporally separated splits, we evaluate standard ML and deep-learning detectors across increasing train--test time gaps. Results show clear temporal degradation, while multimodal fusion outperforms the best single modality across long-term temporal gaps. Cross-modal agreement also declines over time, suggesting that drift affects both individual feature spaces and the consistency among modalities. We further analyze modality-specific drift, malware-family evolution, and temporal changes in model explanations. We publicly release McNdroid, benchmark splits, and code to support reproducible research on temporal generalization and robust multimodal learning in security-critical, non-stationary settings.
| Comments: | 28 pages, 14 figures, 14 tables |
| Subjects: | Cryptography and Security (cs.CR); Machine Learning (cs.LG) |
| Cite as: | arXiv:2605.06894 [cs.CR] |
| (or arXiv:2605.06894v1 [cs.CR] for this version) | |
| https://doi.org/10.48550/arXiv.2605.06894 arXiv-issued DOI via DataCite (pending registration) |
From: Mohammad Saidur Rahman [view email]
[v1]
Thu, 7 May 2026 19:53:24 UTC (31,387 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。