The efficiency of the profiling side-channel analysis can be significantly improved with machine learning techniques. Although powerful, a fundamental machine learning limitation of being data-hungry received little attention in the side-channel community. In practice, the maximum number of leakage traces that evaluators/attackers can obtain is constrained by the scheme requirements or the limited accessibility of the target. Even worse, various countermeasures in modern devices increase the conditions on the profiling size to break the target. This work demonstrates a practical approach to dealing with the lack of profiling traces. Instead of learning from a one-hot encoded label, transferring the labels to their distribution can significantly speed up the convergence of guessing entropy. By studying the relationship between all possible key candidates, we propose a new metric, denoted Label Correlation (LC), to evaluate the generalization ability of the profiling model. We validate LC with two common use cases: early stopping and network architecture search, and the results indicate its superior performance.

Hyperparameter tuning represents one of the main challenges in deep learning-based profiling side-channel analysis. For each different side-channel dataset, the typical procedure to find a profiling model is applying hyperparameter tuning from scratch. The main reason is that side-channel measurements from various targets contain different underlying leakage distributions. Consequently, the same profiling model hyperparameters are usually not equally efficient for other targets. This paper considers autoencoders for dimensionality reduction to verify if encoded datasets from different targets enable the portability of profiling models and architectures. Successful portability reduces the hyperparameter tuning efforts as profiling model tuning is eliminated for the new dataset, and tuning autoencoders is simpler. We first search for the best autoencoder for each dataset and the best profiling model when the encoded dataset becomes the training set. Our results show no significant difference in tuning efforts using original and encoded traces, meaning that encoded data reliably represents the original data. Next, we verify how portable is the best profiling model among different datasets. Our results show that tuning autoencoders enables and improves portability while reducing the effort in hyperparameter search for profiling models. Lastly, we present a transfer learning case where dimensionality reduction might be necessary if the model is tuned for a dataset with fewer features than the new dataset. In this case, tuning of the profiling model is eliminated and training time reduced.

Convolutional neural networks (CNNs) offer unrivaled performance in profiling side-channel analysis. This claim is corroborated by numerous results where CNNs break targets protected with masking and hiding countermeasures. One hiding countermeasure commonly investigated in related works is desynchronization (misalignment). The conclusions usually state that CNNs can break desynchronization as they are shift-invariant. This paper investigates that claim in more detail and reveals that the situation is more complex. While CNNs have certain shift-invariance, it is insufficient for commonly encountered scenarios in deep learning-based side-channel analysis. We investigate data augmentation to improve the shift-invariance and, in a more powerful version, ensembles of data augmentation. Our results show that the proposed techniques work very well and improve the attack significantly, even for an order of magnitude.