Paper 2021/1592

The Need for Speed: A Fast Guessing Entropy Calculation for Deep Learning-based SCA

Guilherme Perin and Lichao Wu and Stjepan Picek

Abstract

In recent years, the adoption of deep learning drastically improved profiling side-channel attacks (SCA). Although guessing entropy is a highly informative metric for profiling SCA, it is time-consuming, especially if computed for all epochs during training. This paper shows that guessing entropy can be efficiently computed during training by reducing the number of validation traces. Our solution significantly speeds up the process, impacting hyperparameter search and profiling attack performances. Our fast guessing entropy calculation is up to 16 times faster and results in more hyperparameter tuning experiments, allowing us to find more efficient deep learning models.