Malware Searching Methods at FML-Architecture

• Published in: East-West Design and Test Symposium pp. 1 - 5
• Main Authors: Hahanov, V.I., Saprykin, A.S.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 10.09.2021
• Subjects: cloud-edge computing; Computational modeling; Computer viruses; cyber-physical system; logic-vector analysis; Machine learning; Machine learning algorithms; malware; malware detection; ML-computing; Neural networks; Productivity; sandbox; Search problems; signature
• ISSN: 2472-761X
• DOI: 10.1109/EWDTS52692.2021.9581024

A matrix-logical method for diagnosing malicious components is being developed, which has high performance due to parallel operations. A vector-matrix method for searching malware in CPS is presented, which uses functionality and malware responses to test datasets. All models and algorithms have been verified in the program code and are ready for implementation in the cloud federated machine learning system for malware detection. The algorithms proposed in the study have linear or quadratic computational complexity. Due to the parallel execution of register logical operations, procedures for analyzing matrices or tables outperform their analogs by two to 10 times in performance.