Cryptology ePrint Archive: Report 2022/165

PAC Learnability of iPUF Variants

Durba Chatterjee and Debdeep Mukhopadhyay and Aritra Hazra

Abstract: Interpose PUF~(iPUF) is a strong PUF construction that was shown to be vulnerable against empirical machine learning as well as PAC learning attacks. In this work, we extend the PAC Learning results of Interpose PUF to prove that the variants of iPUF are also learnable in the PAC model under the Linear Threshold Function representation class.

Category / Keywords: foundations / Physically Unclonable Funcitons, PAC Learning, Boolean Functions

Date: received 13 Feb 2022, last revised 14 Feb 2022

Contact author: durba chatterjee94 at gmail com

Available format(s): PDF | BibTeX Citation

Version: 20220220:200529 (All versions of this report)

Short URL: ia.cr/2022/165