AI for Computer Security

We unleash the power of artificial intelligence to ease the worsening computer security problem like information leakage and side channel attacks.

  • J. Cui, X. Yang*, M. Luo*, G. Lee*, et. al., “MACTA: A Multi-agent Reinforcement Learning Approach for Cache Timing Attacks and Detection”, accepted to International Conference on Learning Representation (ICLR), 2023. [pdf] (* Equal contributions.),[code]

  • M. Luo*, W. Xiong*, et. al., “AutoCAT: Reinforcement Learning for Automated Exploration of Cache Timing-Channel Attacks”, accepted to IEEE International Symposium on High Performance Computer Architecture (HPCA), 2023. [pdf][code](* Equal contributions.)(Artifact Evaluated, Functional and Result Reproduced).

AI System Security

Emerging AI systems like large language models and retrival augmented generation create new security concerns that we must address to safely use these systems.

Hardware Security

Computations ultimately needs to be executed on computer hardware, which are inherently vulnerable. We study how these hardware vulnerabilities can affect software and applications.

  • M. Luo, A. C. Myers, G. E. Suh, “Stealthy Tracking of Autonomous Vehicles with Cache Side Channels”, in 29th USENIX Security Symposium, 2020, pp.859-876 [pdf] [slides][talk] (Shortlisted for Top Picks in Hardware and Embedded Security 2022.)

  • A. Cathis, M. Luo, M. Tiwari, A. Gerstlauer, “Lifecycle-Aware Power-Based Malware Detection”, accepted to HOST, 2025. [pdf]

  • X. Jiao, M. Luo, J. H. Lin, R. K. Gupta, “An Assessment of Vulnerability of Hardware Neural Networks to Dynamic Voltage and Temperatrue Variations”, in Internaional Conference on Computer-Aided Design (ICCAD), 2017, 940-950. [pdf]