2021/02/15 by Thomas Cochrane, Peter Foster, Cochrane, Thomas +9 · 4 citations
Computer Science · #60L10 #Advanced Malware Detection Techniques #Anomaly Detection Techniques and Applications #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #Network Security and Intrusion Detection
paper · pdf · doi:10.48550/arxiv.2102.07904
openalex publication_date 2021/02/15 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
The development of machine learning algorithms in the cyber security domain\nhas been impeded by the complex, hierarchical, sequential and multimodal nature\nof the data involved. In this paper we introduce the notion of a streaming tree\nas a generic data structure encompassing a large portion of real-world cyber\nsecurity data. Starting from host-based event logs we represent computer\nprocesses as streaming trees that evolve in continuous time. Leveraging the\nproperties of the signature kernel, a machine learning tool that recently\nemerged as a leading technology for learning with complex sequences of data, we\ndevelop the SK-Tree algorithm. SK-Tree is a supervised learning method for\nsystematic malware detection on streaming trees that is robust to irregular\nsampling and high dimensionality of the underlying streams. We demonstrate the\neffectiveness of SK-Tree to detect malicious events on a portion of the\npublicly available DARPA OpTC dataset, achieving an AUROC score of 98%.\n