Profiled side-channel attacks consist of several steps one needs to take. An important, but sometimes ignored, step is a selection of the points of interest (features) within side-channel measurement traces. A large majority of the related works start the analyses with an assumption that the features are preselected. Contrary to this assumption, here, we concentrate on the feature selection step. We investigate how advanced feature selection techniques stemming from the machine learning domain can be used to improve the attack efficiency. To this end, we provide a systematic evaluation of the methods of interest. The experiments are performed on several real-world data sets containing software and hardware implementations of AES, including the random delay countermeasure. Our results show that wrapper and hybrid feature selection methods perform extremely well over a wide range of test scenarios and a number of features selected. We emphasize L1 regularization (wrapper approach) and linear support vector machine (SVM) with recursive feature elimination used after chi-square filter (Hybrid approach) that performs well in both accuracy and guessing entropy. Finally, we show that the use of appropriate feature selection techniques is more important for an attack on the high-noise data sets, including those with countermeasures, than on the low-noise ones.

@en

The profiled side-channel analysis represents the most powerful category of side-channel attacks. In this context, the security evaluator (i.e., attacker) gains access to a profiling device to build a precise model which is used to attack another device in the attacking phase. Mostly, it is assumed that the attacker has significant capabilities in the profiling phase, whereas the attacking phase is very restricted. We step away from this assumption and consider an attacker restricted in the profiling phase, while the attacking phase is less limited. We propose the concept of semi-supervised learning for side-channel analysis, where the attacker uses a small number of labeled measurements from the profiling phase as well as the unlabeled measurements from the attacking phase to build a more reliable model. Our results show that the semi-supervised concept significantly helps the template attack (TA) and its pooled version (TA_p). More specifically, for low noise scenario, the results for machine learning techniques and TA are often improved when only a small number of measurements is available in the profiling phase, while there is no significant difference in scenarios where the supervised set is large enough for reliable classification. For high noise scenario, TA_p and multilayer perceptron results are improved for the majority of inspected dataset sizes, while for high noise scenario with added countermeasures, we show a small improvement for TA_p, Naive Bayes and multilayer perceptron approaches for most inspected dataset sizes. Current results go in favor of using semi-supervised learning, especially self-training approach, in side-channel attacks.

@en

We concentrate on machine learning techniques used for profiled side-channel analysis in the presence of imbalanced data. Such scenarios are realistic and often occurring, for instance in the Hamming weight or Hamming distance leakage models. In order to deal with the imbalanced data, we use various balancing techniques and we show that most of them help in mounting successful attacks when the data is highly imbalanced. Especially, the results with the SMOTE technique are encouraging, since we observe some scenarios where it reduces the number of necessary measurements more than 8 times. Next, we provide extensive results on comparison of machine learning and side-channel metrics, where we show that machine learning metrics (and especially accuracy as the most often used one) can be extremely deceptive. This finding opens a need to revisit the previous works and their results in order to properly assess the performance of machine learning in side-channel analysis. @en