On the Use of Machine Learning to Detect Cybersecurity Incidents in Critical Systems

Event
  • Duration: 19 mins
  • Publication date: 02 Nov 2018

Abstract

We begin by reviewing what AI/ML is, what its key and established techniques are and overview typical applications with a focus on Critical Systems or similar such as Industrial Control Systems. We then compare existing literature and projects to produce a survey on their effectiveness in detecting, and acting on, cybersecurity anomalies. While pointing gaps and limitations, we draft a safety and cybersecurity architecture that leverages its power to enhance the current systems. Finally, we discuss at some length the typical development and integration process and compare with the same for safety critical systems.

Keywords:
Internet safety