AI-driven Detection of Cybersecurity-related Patterns

Abstract

The considerable increase in the number of attacks and online violations in recent years highlighted how security is more often considered a feature than a necessity. Indeed, most of the solutions proposed over time have often favored efficiency, promptness, and quality of service, to the detriment of data and information security. Not surprisingly, threat actors have exploited these gaps in systems and managed to score attacks to undermine their security, thus getting their hands on information that was thought to remain secret. This opportunity gives excessive power to malicious users who, depending on their intentions, have the potential to cause massive damage. To make things worse, attackers are relying on Artificial intelligence technology to automatically launch attack with an ever-increasing impact. To deal with these new challenges, researchers and organizations started fighting fire with fire by investing time and money to explore Artificial Intelligence's employment in their research and cyber risk operations. Artificial Intelligence's adoption allows the new weaknesses, threats, and exploits to be more promptly identified and analyzed to help relieve further attacks. During my Ph.D. I investigated how the adoption of Artificial Intelligence technologies would help identify, detect, and automate cybersecurity-related tasks, for defensive purposes. In particular, I decided to investigate the AI-driven detection and recognition of cybersecurity-related patterns taking into account five domain: the wireless keyboard domain, the digital forensic domain, the satellite authentication domain, the cryptojacking domain, and the Dark Web domain, respectively.

Date of Award	2021
Original language	American English
Awarding Institution	HBKU College of Science and Engineering