PhD in Computer Science and Engineering - Secure, Trustworthy, and Reliable Systems in Ann Arbor United States | University of Michigan Ann Arbor

The ongoing miniaturization of silicon manufacturing processes has introduced major reliability issues, now emerging as a primary design obstacle. Design robustness faces threats from both complexity hurdles - where massive designs resist full verification - and physical limitations like chip degradation and transient faults that compromise proper system function, alongside hardware-exploitable security vulnerabilities. CSE researchers in this field are tackling these critical issues through coordinated research approaches, spanning from immediate reliability enhancement methods to boost current chip quality, to future-oriented solutions for identifying, recovering from, and fixing system failures. Furthermore, ensuring the absolute functional accuracy of digital chips, especially sophisticated processors, is crucial for creating secure and dependable systems. Yet this objective remains unmet today, as shown by extensive bug reports for commercial processors documenting design flaws missed during verification. To overcome verification difficulties, researchers are developing self-monitoring systems that can identify and rectify their own errors. They're pioneering update methods to fix these undetected bugs on-site, effectively giving hardware software-like adaptability. Additionally, they're studying economical methods for real-time computation validation, specifically approaches guaranteed to block erroneous outputs.