We study security risks in AI-powered software systems and develop techniques to analyze, evaluate, and defend against attacks on intelligent agents and AI-assisted applications.
We study binary analysis and reverse engineering techniques to understand program behavior, recover program semantics, and support security analysis of real-world software.
This research is about to protect the operating system kernel against attacks by ensuring the integrity, confidentiality, and availability of kernel.
This project is about to explore techiniques to find vulnerabilities in software automatically,including fuzzing, symbolic execution and so on.
Compilers are important and complex, in this project, we study various security issues introduced by compilers.
We explore how AI can assist program analysis, software testing, vulnerability detection, and security reasoning.