Being Patient and Persistent: Optimizing An Early Stopping Strategy for Deep Learning in Profiled Attacks

The absence of an algorithm that effectively monitors the deep learning models used in side-channel attacks increases the difficulty of a security evaluation. If an attack is unsuccessful, that could be due to multiple reasons. It can be that we are indeed dealing with a resistant implementation, but it is possible that the deep learning model used is faulty. In this contribution, we formalize two conditions, persistence and patience, for a deep learning model to be optimal and we propose an early stopping algorithm that reliably recognizes the model's optimal state during training. The novelty of our solution is in an efficient implementation of guessing entropy estimation as a success metric used to measure the strength of a side-channel adversary. As a result, the model which uses our strategy for learning converges with fewer traces than other known methods.