Enhancing Real-Time Hardware-based Mobile Malware Detection through Multiple Instance Learning

A novel Multiple Instance Learning (MIL) formulation that accurately represents localized malware behaviors in segmented hardware telemetry time-series, improving the precision of real-time malware detection.