On the performance of non‐profiled side channel attacks based on deep learning techniques

Do, Ngoc-Tuan; Hoang, Van‐Phuc; Doan, Van‐Sang; Pham, Cong‐Kha

doi:10.1049/ise2.12102

Cited by 9 publications

(4 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A technique for conducting sensitivity analysis known as gradient visualisation was used by Masure et al in order to locate the areas where information was being leaked [15]. Van der Valk and Picek improved upon bias-variance decomposition and introduced GE bias-variance decomposition [16] in…”

Section: International Journal Of Professional Studiesmentioning

confidence: 99%

Design of Deep Learning Technique Based Side Channel Attack Analysis for System on Chips

Fattah

2024

IJPS

View full text Add to dashboard Cite

The effectiveness of deep learning techniques has increased in recent years, making it possible to assess side channel attacks on System-on-Chips (SoCs). Specifically, this is due to the fact that they provide a sophisticated method to capitalise on the unintended loss of information that occurs during cryptographic procedures. In this particular scenario, it is very necessary to collect information on side channels, such as power consumption or electromagnetic emissions. There is a procedure called as preprocessing that involves cleaning and modifying the raw data in order to make it more acceptable for input into neural networks. The specifics of the side channel information are what define whether or not a deep learning architecture, such as Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM) networks, or other specialist structures, should be used. Following this, the architecture of the model is painstakingly created, consisting of layers, units, and activation functions, with the purpose of effectively collecting and deciphering the intricate patterns that are associated with the processing of sensitive data on the SOC.As part of this study, we presented a CNN-dependent profiled SCA assault that was directed against an AES cypher that was operating on a SoC. During the phase of setup, we focus on the discharge from a minor capacitor that is linked with the principal line of power supply. This capacitor produces a distinct signal, and we do not capture electromagnetic emission of chip surface. When it comes to the matching and profiling phases, CNN is then used. In light of the fact that the most recent AES round did not uncover any leaks, it is recommended that a novel approach be used in order to retrieve the keys from the first round, Through the use of AES algorithm observations, we were able to develop a differential equation that included the selected intermediate value and basic text. This equation in the end resulted in possibility of advantageous for incorrect value.

show abstract

Section: International Journal Of Professional Studiesmentioning

confidence: 99%

Design of Deep Learning Technique Based Side Channel Attack Analysis for System on Chips

Fattah

2024

IJPS

View full text Add to dashboard Cite

show abstract

“…According to Martinasek et al [13], it was shown that MLP can crack naïve AES. In addition, R. Gilmore et.…”

Section: Review Of Workmentioning

confidence: 99%

Design of Deep Learning Techniques for Side-Channel Attacks on Masked 128-bit AES Implementations

Saeb Nahi,

Fattah,

Mutashar

et al. 2024

KJCS

View full text Add to dashboard Cite

Researchers are exploring the use of convolutional neural networks (CNNs) in side-channel attacks to understand the weaknesses in cryptographic implementation. CNNs can learn hierarchical characteristics automatically from electromagnetic radiation or power usage during cryptographic processes. Researchers train CNNs on side-channel data to extract meaningful representations and deduce secret keys. Deep learning algorithms are helpful in evaluating the security of embedded systems, and CNNs are a feasible paradigm for profiling side-channel analysis attacks. In this paper, it has been introduced a VGG (Visual Geometry Group)-Net architecture, which is a typical deep convolutional neural network design with numerous layers. It uses the ASCAD dataset to conduct experiments. They found that VGG-Net architecture Side Channel Attacks (SCA) provides better results than the previously optimized CNN model by significantly reducing the number of side-channel traces required for successful attacks on desynchronized datasets. The researchers also discovered that synchronous traces serve as the pre-training source for VGG-Net architecture, functioning successfully in terms of jittering with minimal fine-adjusting after training

show abstract

“…Then, several works take into account different machine learning approaches to carry out tuning of hyperparameter by randomized or grid searching. The authors of [19] undertake an experimental assessment of various hyper-parameters concerning CNNs using ASCAD repository. To create deep learning model ensembles, Perin et al used a random search inside pre-specified ranges [20].…”

Section: Literature Reviewmentioning

confidence: 99%

Adaptive Optimization of Deep Learning Models on AES based Large Side Channel Attack Data

Muhammad,

Alzuabidi,

Ahmed

et al. 2024

KJCS

View full text Add to dashboard Cite

Deep learning-based side-channel analysis is an efficient and suitable technique for profiling side-channel attacks. In order to obtain the better performance, it is highly necessary to analyze an in-depth training stage in which the optimization of relevant hyperparameters should be a vital process. During the training phase, hyperparameters that are connected to the architecture of the neural network are often selected; however, hyperparameters that impact the training process are to be effectively analyzed. This was represented by an optimized hyperparammeter that consists of considerable impact on attacking behaviour, which is the primary focus of our research. Our research has shown that even while the popular optimizers Adam and RMSprop are capable of delivering satisfactory outcomes, they are also tend to being overfit. Hence, it is necessary to use condensed training periods, simple profiling models, and explicit regularization in order to avoid this problem. On the other hand, the performance of optimizers of the SGD type is only satisfactory when momentum is used which results in slower convergence and less overfit. In conclusion, the research results provide a better use of Adagrad in the cases of longer training datasets or big profiling models.

show abstract

On the performance of non‐profiled side channel attacks based on deep learning techniques

Cited by 9 publications

References 22 publications

Design of Deep Learning Technique Based Side Channel Attack Analysis for System on Chips

Design of Deep Learning Technique Based Side Channel Attack Analysis for System on Chips

Design of Deep Learning Techniques for Side-Channel Attacks on Masked 128-bit AES Implementations

Adaptive Optimization of Deep Learning Models on AES based Large Side Channel Attack Data

Contact Info

Product

Resources

About