Dynamic Entropic Signatures for Ransomware Detection: A Novel Computational Framework
Daniel Azzaman, David Spyridon, Henry White, Matthew Kassiani · 2024
The growing complexity and frequency of encryption-based cyberattacks have significantly heightened the urgency for advanced detection mechanisms capable of adapting to sophisticated adversarial tactics. Current methods often fall short in addressing the rapid evolution of ransomware techniques, which exploit system vulnerabilities and evade conventional security protocols. The introduction of a novel computational framework, Dynamic Entropic Signatures (DES), offers an innovative approach to ransomware detection through the dynamic analysis of entropy variations within system activities. By capturing and analyzing the temporal entropy patterns associated with malicious encryption processes, the framework achieves superior accuracy, precision, and recall compared to traditional detection methodologies. Extensive empirical evaluations demonstrated its robust adaptability across diverse ransomware samples and file types while maintaining computational efficiency and minimizing false positives. The findings reinforce the potential of DES to transform the cybersecurity landscape, providing a scalable, real-time solution for defending critical digital infrastructures against increasingly sophisticated threats.