Xingliang Yuan

Xingliang contributes practical research on source-code vulnerability detection and the protection of AI models in mobile software. SAFE’s public code gives researchers a basis for examining the approach, while THEMIS addresses ownership protection for mobile model developers and third parties such as app stores. Evaluation on real-world apps connects the research to deployment challenges; cybersecurity course leadership also supports education in the field.

Xingliang Yuan is an Associate Professor and ARC Future Fellow at the University of Melbourne, serving as Course Director for the Master of Cybersecurity. Before moving to Melbourne in June 2024, Xingliang was a faculty member at Monash University from 2017 to 2024. Research areas span secure networked systems, encrypted data processing, and making machine learning safer to deploy.

Key collaborative projects include SAFE, an approach to identifying security weaknesses in source code with a public implementation, and THEMIS, a system for protecting AI models distributed inside mobile apps. Xingliang investigated ways to embed watermarks into models post-deployment, addressing the risk of model theft when running machine learning locally on devices. An evaluation across 403 real-world Google Play apps demonstrated an 81.14% success rate.

Alongside research, Xingliang contributes to cybersecurity education and academic exchange through course leadership and co-organising events, including the Workshop on Secure Collaborative Computing at Scale. These initiatives help software practitioners and researchers investigate vulnerabilities and secure critical digital assets.