Efficient Nonprofiled Side-Channel Attack Using Multi-Output Classification Neural Network

Abstract

Differential Deep Learning Analysis (DDLA) is the first deep learning based non-profiled side-channel attack (SCA) on embedded systems. However, DDLA requires many training processes to distinguish the correct key. In this letter, we introduce a non-profiled SCA technique using multi-output classification to mitigate the aforementioned issue. Specifically, a multi-output multi-layer perceptron and a multi-output convolutional neural network are introduced against various SCA protected schemes, such as masking, noise generation, and trace de-synchronization countermeasures. The experimental results on different power side channel datasets have clarified that our model performs the attack up to 9 and 30 times faster than DDLA in the case of masking and de-synchronization countermeasures, respectively. In addition, regarding combined masking and noise generation countermeasure, our proposed model achieves a higher success rate of at least 20% in the cases of the standard deviation equal to 1.0 and 1.5.