Optimal Tradeoff of Secure PUF-based Authentication

We consider a problem of secure PUF-based authentication under a privacy constraint on the PUF responses. Given a randomly chosen challenge, an adversary who has access to the (publicly) stored helper data and side information correlated with the challenge -response pairs may try to mimic the corresponding PUF response in order to get authenticated. We characterize the optimal tradeoff between rate of the compressed helper data, information leakage rate of the PUF response at the adversary, and the adversary's maximum false acceptance probability exponent.. The result captures a tradeoff between authentication performance and security aspects of PUF-based authentication such as authentication security and unpredictability of PUF when used in authentication applications.