CODES 2026 Paper Accepted: Physically Coupled Layered PUF for Machine Learning Resilience

I am pleased to share that our collaborative paper, “PCL-PUF: Physically Coupled Layered PUF for Machine Learning Resilience,” has been accepted to the ACM/IEEE International Conference on Hardware/Software Codesign and System Synthesis (CODES 2026).

Physical Unclonable Functions (PUFs) are widely used as lightweight hardware security primitives for authentication and cryptographic key generation. However, recent advances in machine learning have made many existing strong PUF designs vulnerable to modeling attacks, motivating the development of more resilient architectures.

In this work, we propose PCL-PUF, a novel architecture that physically couples delay-based PUFs with non-volatile-memory (NVM)-based PUFs. By dynamically modulating the physical behavior of the lower layer, the design significantly increases the complexity of machine-learning-based modeling while maintaining desirable PUF characteristics.

Comprehensive evaluation demonstrates resilience against four major categories of modeling attacks: logistic regression, reliability, splitting, and neural-network attacks, showing substantial improvements over comparable state-of-the-art PUF designs.

Congratulations to all co-authors on this achievement!

Authors: Mohamed Alsharkawy, Hassan Nassar, Jeferson Gonzalez, Kuan-Hsun Chen, and Jörg Henkel.