The Owhiti Cybersecurity research group are hosting a seminar featuring four presentations on artificial intelligence applications in cybersecurity.

Professor M. Ali Babar (University of Adelaide) will present findings on the empirical evaluation of Large Language Models for software security. His talk covers the technological, organizational, and socio-psychological aspects of leveraging LLMs for security-by-design paradigms, addressing methodological considerations and challenges in conducting evaluation studies of LLM-based security tools.

Dr. Triet Le (University of Adelaide) will discuss data-centric solutions for AI-powered software vulnerability prediction. Based on large-scale empirical studies across hundreds of real-world projects, the presentation examines how identifying latent vulnerabilities, automatic labeling, noise-reduction strategies, and contextual data augmentation can improve the performance and reliability of vulnerability prediction models.

Dr. Etienne Borde (University of Canterbury, New Zealand) will present a framework for optimizing Moving Target Defense (MTD) strategies using reinforcement learning and multi-agent simulation. Based on work published at ARES 2025, the talk introduces a reward function that models trade-offs between security and system availability, and discusses challenges including scalability and realism of attacker behaviors.

Dr. Mengmeng Ge (Monash University) will present three intrusion prevention mechanisms for IoT networks: graphical security modeling approaches for assessing system vulnerabilities, cyber deception techniques, and moving target defense strategies to protect resource-constrained IoT devices from cyberattacks.