1. How to Fool a Black Box Machine Learning Based Side-Channel Security Evaluation
- Author
-
François-Xavier Standaert, Gaëtan Cassiers, Charles-Henry Bertrand Van Ouytsel, Olivier Bronchain, UCL - SST/ICTM/ELEN - Pôle en ingénierie électrique, and UCL - SST/ICTM - Institute of Information and Communication Technologies, Electronics and Applied Mathematics
- Subjects
Computer Networks and Communications ,Computer science ,0102 computer and information sciences ,02 engineering and technology ,Machine learning security ,Black box side channel security evaluation ,Machine learning ,computer.software_genre ,01 natural sciences ,Machine Learning ,0202 electrical engineering, electronic engineering, information engineering ,Side channel attack ,Implementation ,Security evaluation ,Black box (phreaking) ,business.industry ,Applied Mathematics ,Deep learning ,020206 networking & telecommunications ,Adversary ,Computational Theory and Mathematics ,010201 computation theory & mathematics ,Key (cryptography) ,Artificial intelligence ,business ,computer - Abstract
Machine learning and deep learning algorithms are increasingly considered as potential candidates to perform black box side-channel security evaluations. Inspired by the literature on machine learning security, we put forward that it is easy to conceive implementations for which such black box security evaluations will incorrectly conclude that recovering the key is difficult, while an informed evaluator / adversary will reach the opposite conclusion (i.e., that the device is insecure given the amount of measurements available).
- Published
- 2021