École Nationale Supérieure de
Techniques Avancées

Abstract

The AES standard is widely used for encryption, its algorithm is public because the security of this standard is based on the secret key that is only known by the sender and recipient of the information, so its security is compromised if an attacker manages to obtain it. side-channel attacks try to recover this key through existing information leaks in the system that implements the AES, these leaks can be obtained through energy consumption, electromagnetic radiation, or any other factor that contains a relation between the secret key and this. The DPA and CPA attacks are part of the side-channel attacks, these use statistics to recover the secret key and its scope has been amply studied, in this research is raised the possibility of using machine learning to obtain the complete secret key using fewer traces than those required for DPA and CPA attacks with a single model.