EPSRC DTP studentship in Cyber Security Analytics: Automated Cyber Defence (with Airbus)

We live in a world of ever-increasing digital interconnectivity and are rapidly moving towards pervasive autonomous systems. Recent widespread cyber attacks such as WannaCry demonstrate the devastating impact an out-of-control attack can have on networked computer systems. While this example could have been prevented by keeping systems up to date, we are inevitably going to see more automation of attacks that are intelligent and difficult to defend against.

The PhD studentship in Automated Cyber Defence will build on the long standing collaboration between Cardiff University & Airbus, and our recent ground-breaking research into the early-detection of cyber attacks - seeing ransomware detected within 4 seconds, and reducing file system damage by over 80%. The PhD will seek to lay foundations in the development of "game playing" cyber attack and defence methods, to pick up the early warning signs of an attack on enterprise networks and "learn how to defend" by selecting appropriate automated defence mechanisms, depending on the context of the attack. Core to this proposition will be innovation in representing attack and defence tactics, and artificial logic and intelligence around how to select a suitable automated response to imminent attacks. We are seeking an open minded, creative individual to join the team and develop world class research outcomes with industrial applications. You will join the ESPRC DTP Hub in Cyber Security Analytics at Cardiff University, becoming part of an interdisciplinary cohort of students studying the human and algorithmic aspects of AI in the context of cybersecurity.

Objectives:

i) develop methods to automatically vulnerability-test an enterprise network and orchestrate a set of attacks against these vulnerabilities

ii) develop methods to learn how to strategically defend the network and its devices against a range of increasingly complex attacks in real-time in a responsible, auditable and explainable...