Deep Learning-based Side-channel Analysis: From Concepts to Practice

Abstract

This tutorial provides a practical introduction to deep learning-based physical side-channel analysis. Starting from the fundamentals of side-channel leakage and profiling attacks, we will explain how neural networks can be used to recover secret information from cryptographic implementations, including protected ones.

The tutorial combines conceptual explanations with hands-on exercises. Participants will work through the main steps of a deep learning-based attack pipeline: understanding trace data, preparing datasets, selecting and training neural network architectures, evaluating attack performance, and interpreting results. Along the way, we will highlight state-of-the-art achievements, common pitfalls, and open research challenges.

By the end of the tutorial, participants will have a clear understanding of how deep learning-based side-channel attacks are designed and evaluated, as well as where the field is heading next.

Speaker