Machine-Learning-Aided Side-Channel Analysis in Post-Quantum Cryptography

National and international asymmetric crypto standards on public-key encryption and digital signatures will soon shift to post-quantum ones that can survive in the coming quantum era. This transition introduces new challenges in achieving physical security against adversaries exploiting various side channels such as timing, power consumption, and radiations. The involvement of machine learning in side-channel analysis hardens this effort.
This two-year project is aimed at designing novel profiled side-channel attacks to the state-of-the-art post-quantum crypto implementations, with the help of machine learning algorithms. We also plan to study the methodology for safe implementation against the designed attacks.

The research is funded by the Crafoord Foundation.