2026/05/07 by Md Mahmuduzzaman Kamol, Jesus Lopez, Saeefa Rubaiyet Nowmi +5 · 1 voice
Computer Science · #Advanced Malware Detection Techniques #Android (operating system) #Android malware #Benchmark (surveying) #Code (set theory) #Data Stream Mining Techniques #Exploit #Inference #Malware #Robustness (evolution) #Sandbox (software development) #Software Testing and Debugging Techniques #cs.CR #cs.LG
paper · pdf · doi:10.48550/arxiv.2605.06894
openalex publication_date 2026/05/07 · arxiv published 2026/05/07 · arxiv updated 2026/05/07 · openalex created_date 2026/05/12 · openalex updated_date 2026/07/28
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.