Lecture 6 - Deep Learning-based Side-Channel Analysis (Guest lecture)


In this guest lecture, given by Stjepan Picek from Radboud University, we turn our attention to implementation attacks on cryptosystems, focusing in particular on Side-Channel Analysis (SCA). SCA exploits physical leakages in the implementation (for instance power, timings, temperature, and so on) to infer secret information during the execution of a cipher. We also see how deep learning nowadays represents the state of the art in side-channel analysis, especially in the area of profiling attacks.

Topics covered:

  • Implementation attacks
  • Side-channel attacks
  • Profiling attacks
  • Deep learning-based profiling attacks
  • Challenges in deep learning-based SCA

Guest Lecturer Short Bio:

Stjepan Picek is an associate professor at Radboud University, The Netherlands. His research interests involve security and cryptography, machine learning, and evolutionary computation. Prior to the associate professor position, Stjepan was an assistant professor at TU Delft, and a postdoctoral researcher at MIT, USA and KU Leuven, Belgium. Stjepan finished his double doctorate in 2015 with a thesis on cryptology and evolutionary computation techniques. Up to now, Stjepan has given more than 40 invited talks and published more than 150 refereed papers. He is a program committee member and reviewer for a number of conferences and journals, and a member of several professional societies.

Reading Material

Survey on Deep Learning-based Side-Channel Analysis:

See also the references mentioned in the slides.

Lecture Recording