Cryptology ePrint Archive: Report 2021/328

Mixup Data Augmentation for Deep Learning Side-Channel Attacks

Karim M. Abdellatif

Abstract: Following the current direction in Deep Learning (DL), more recent papers have started to pay attention to the efficiency of DL in breaking cryptographic implementations. Several works focus on techniques to boost the efficiency of existing architectures by data augmentation, regularization, etc. In this work, we investigate using mixup data augmentation \cite{zhang2017mixup} in order to improve the efficiency of DL-based Side-Channel Attacks (SCAs). We validated the soundness of the mixup on real traces collected from the ChipWhisperer board \cite{cw} and from the ASCAD database \cite{benadjila2020deep}. The obtained results have proven that using mixup data augmentation decreases the number of measurements needed to reveal the secret key compared to the non-augmented case.

Category / Keywords: Deep Learning, Side-Channel Attacks, Data Augmentation, Mixup and AES

Date: received 11 Mar 2021

Contact author: karim m abdellatif at gmail com

Available format(s): PDF | BibTeX Citation

Version: 20210312:072439 (All versions of this report)

Short URL: ia.cr/2021/328


[ Cryptology ePrint archive ]