| Manuel San Pedro |  |
 |
Date de l'exposé : 12 mai 2017, 10h30-11h30, salle Petri/Turing
Side Channel from a Machine Learning point of view
Side Channel Analysis aims at exploiting a physical leak of a sensitive operation in order to expose a secret.Recent works in the area showed that the use of Machine Learning techniques could improve the analysis at almost every single step of the attack.
This talk aim at presenting how a side-channel attacker could benefit from this discipline to tackle down his analysis by reducing his problems to well known Machine Learning instance. Signal acquisition, resynchronization, features selection, profiled attacks, high-order, ..., could all benefit from objects inherited from Machine Learning. In particular, the emergence of Deep Learning, made possible by advances in hardware (GPU mostly), made possible to build evolved neural network that are able to solve most of these problems.
