Applying Generative Machine Learning to Intrusion Detection: A Systematic Mapping Study and Review

Applying Generative Machine Learning to Intrusion Detection: A Systematic Mapping Study and Review

June 2024 | JAMES HALVORSEN, CLEMENTE IZURIETA, HAIPENG CAI, ASSEFAW GEBREMEDHIN
The article "Applying Generative Machine Learning to Intrusion Detection: A Systematic Mapping Study and Review" by James Halvorsen, Clemente Izurieta, Haipeng Cai, and Assefaw Gebremedhin explores the application of Generative Machine Learning Models (GMLMs) in intrusion detection systems (IDSs). The authors conduct a systematic mapping study and provide a detailed review of the literature, focusing on three main application areas: (1) GMLMs for assisting with penetration testing, (2) GMLMs for supplementing IDS datasets, and (3) GMLMs as IDSs. They discuss the challenges and opportunities in using GMLMs, such as the lack of quality training data and high false-positive rates, and how GMLMs can help overcome these issues. The article also reviews evaluation metrics used in different application areas and highlights the need for standardized metrics. The authors synthesize the reviewed works, identify cross-cutting themes, and suggest future research directions, emphasizing the importance of creating realistic synthetic data for IDSs and addressing the ethical concerns of offensive utility.The article "Applying Generative Machine Learning to Intrusion Detection: A Systematic Mapping Study and Review" by James Halvorsen, Clemente Izurieta, Haipeng Cai, and Assefaw Gebremedhin explores the application of Generative Machine Learning Models (GMLMs) in intrusion detection systems (IDSs). The authors conduct a systematic mapping study and provide a detailed review of the literature, focusing on three main application areas: (1) GMLMs for assisting with penetration testing, (2) GMLMs for supplementing IDS datasets, and (3) GMLMs as IDSs. They discuss the challenges and opportunities in using GMLMs, such as the lack of quality training data and high false-positive rates, and how GMLMs can help overcome these issues. The article also reviews evaluation metrics used in different application areas and highlights the need for standardized metrics. The authors synthesize the reviewed works, identify cross-cutting themes, and suggest future research directions, emphasizing the importance of creating realistic synthetic data for IDSs and addressing the ethical concerns of offensive utility.
Reach us at info@study.space