This paper presents an interconnect obfuscation methodology at the Register-Transfer Level (RTL) using Switch Boxes (SBs) constructed of Polymorphic Transistors to protect integrated circuit designs from unauthorized use and illegal overproduction. Security-aware high-level synthesis algorithms are also introduced to strategically assign RTL interconnects to functional units in a way that corrupts multiple outputs when the polymorphic SBs are incorrectly unlocked.
ChatFuzz leverages large language models and reinforcement learning to generate complex, interdependent, and pseudo-random instruction sequences that significantly improve hardware coverage and vulnerability detection in modern processors.
The WESEE attack abuses the VMM Communication Exception (#VC) in AMD SEV-SNP to compromise the confidentiality and integrity of guest VMs by injecting malicious #VCs that induce arbitrary register and memory read/write operations.
Automated RL framework for HT insertion and detection to address benchmark limitations and human bias.
Hardware fuzzing is crucial for identifying vulnerabilities in complex hardware designs.
DECOR is a randomized algorithm-based method that significantly decreases the correlation between locked circuit netlist and correct key values, enhancing resilience against ML-based attacks.
WhisperFuzz introduces a novel white-box fuzzer with static analysis to detect and locate timing vulnerabilities in processors, addressing deficiencies in existing fuzzers.
The authors propose a novel methodology to verify data-oblivious behavior in hardware using standard property checking techniques, ensuring scalability even to complex out-of-order cores.
The author presents DECOR, a method to enhance logic locking schemes against machine learning attacks by decorrelating the circuit structure from the correct key, significantly reducing key prediction accuracy.