SoK: Deep Learning-based Side-channel Analysis Trends and Challenges

Sengim Karayalcin, Marina Krcek, Stjepan Picek

Abstract: Deep learning-based side-channel analysis (DLSCA) represents a powerful paradigm for running side-channel attacks. DLSCA in a state-of-the-art can break multiple targets with only a single attack trace, requiring minimal feature engineering. As such, DLSCA also represents an extremely active research domain for both industry and academia. At the same time, due to domain activity, it becomes more difficult to understand what the current trends and challenges are. In this systematization of knowledge, we provide a critical outlook on a number of developments in DLSCA in the last year, allowing us to offer concrete suggestions. Moreover, we examine the reproducibility perspective, finding that many works still struggle to provide results that can be used by the community.

External IDs:dblp:journals/iacr/KarayalcinKP25a