{"id":413,"date":"2025-11-17T09:00:29","date_gmt":"2025-11-17T09:00:29","guid":{"rendered":"https:\/\/blogs.kent.ac.uk\/cyber-institute-news\/?p=413"},"modified":"2025-11-17T11:38:46","modified_gmt":"2025-11-17T11:38:46","slug":"icss-researchers-won-distinguished-paper-award-at-autonomouscyber-2025-workshop","status":"publish","type":"post","link":"https:\/\/blogs.kent.ac.uk\/cyber-institute-news\/2025\/11\/17\/icss-researchers-won-distinguished-paper-award-at-autonomouscyber-2025-workshop\/","title":{"rendered":"iCSS researchers won Distinguished Paper Award at AutonomousCyber 2025 Workshop"},"content":{"rendered":"

Two iCSS researchers from the University of Kent\u2019s School of Computing, Joshua Sylvester<\/a> (PhD student) and his supervisor Dr Rog\u00e9rio de Lemos<\/a> (Senior Lecturer and Director of PGR Studies), won the Distinguished Paper Award<\/strong> at the 2nd International Workshop on Autonomous Cybersecurity (AutonomousCyber 2025)<\/a>, co-located with the 30th European Symposium on Research in Computer Security (ESORICS 2025)<\/a>, one of the top-tier cyber security research conferences.<\/p>\n

The abstract of the award-winning paper, titled \u201cKnowledge Retention for Generic Reinforcement Learning Policies in Autonomous Cyber Defence\u201d, is as follows:<\/p>\n

\u201cWithin autonomous cyber defence, building scalable agents that can generalise across attack behaviours is crucial to developing a truly autonomous system. These generic agents are pivotal, as over time, attackers will inevitably change their behaviour, requiring the defence mechanisms to adapt accordingly. Current approaches for generic agents use deep reinforcement learning policies to learn multiple attack behaviours and mitigate them. When a new attack behaviour is introduced, the generic policy is retrained to incorporate this behaviour and not forget previous attack behaviours. In this paper, we propose a novel solution based on a modified version of the Proximal Policy Optimisation (PPO) reinforcement learning algorithm that retains previously acquired knowledge, enabling a scalable and generic framework in which new attack behaviours can be incorporated modularly. The modified PPO algorithm demonstrates a 22.11% performance improvement compared to standard PPO when trained to sequentially learn two distinct attack behaviours. These results show a step towards building more scalable autonomous cyber defence systems capable of incorporating evolving cyber threats.<\/em>\u201d<\/p>\n

The paper\u2019s full citation information is the following:<\/p>\n