Understanding and Taming Deterministic Model Bit Flip Attacks in Deep Neural Networks
This project aims to offer insights into DNN security issues due to hardware-based fault attacks, and explore ways to promote the robustness and security of future deep learning system against such internal adversaries.
Uncovering and Defeating Microarchitecture Attacks in Modern Processors
This project investigates side and covert channels in microarchitecture design of modern processors and aims to design low-cost architectural and system-level defensive approaches to mitigate these attacks.
Architecting Secure-By-Design ReRAM-based Memories
This project will investigate the security implications of ReRAM-based NVMs in computing systems and how to structure secure-by-design ReRAM-based memories.
Towards Secure-By-Design Integration of Emerging Non-Volatile Memory in Future System
This project will design efficient architecture and system level defensive techniques to defeat futuristic adversaries exploiting the discovered information leakage vulnerabilities in NVM-based systems