Deep learning represents a powerful set of techniques for profiling sidechannel analysis. The results in the last few years show that neural network architectures like multilayer perceptron and convolutional neural networks give strong attack performance where it is possible to break targets protected with various countermeasures. Considering that deep learning techniques commonly have a plethora of hyperparameters to tune, it is clear that such top attack results can come with a high price in preparing the attack. This is especially problematic as the side-channel community commonly uses random search or grid search techniques to look for the best hyperparameters.In this paper, we propose to use reinforcement learning to tune the convolutional neural network hyperparameters. In our framework, we investigate the Q-Learning paradigm and develop two reward functions that use side-channel metrics. We mount an investigation on three commonly used datasets and two leakage models where the results show that reinforcement learning can find convolutional neural networks exhibiting top performance while having small numbers of trainable parameters. We note that our approach is automated and can be easily adapted to different datasets. Several of our newly developed architectures outperform the current state-of-the-art results. Finally, we make our source code publicly available.
https://github.com/AISyLab/Reinforcement-Learning-for-SCA
Abstract. Expanding of karst rocky desertification (RD) area in southwestern China is strangling the sustainable development of local agricultural economy. It is important to evaluate the soil fertility at RD regions for the sustainable management of karst lands. The changes in 19 different soil fertilityrelated variables along a gradient of karst rocky desertification were investigated in five different counties belonging to the central Hunan province in China. We used principal component analysis method to calculate the soil data matrix and obtained a standardized integrate soil fertility (ISF) indicator to reflect RD grades. The results showed that the succession of RD had different impacts on soil fertility indicators. The changing trend of total organic carbon (TOC), total nitrogen (TN), available phosphorus, microbial biomass carbon (MBC), and microbial biomass nitrogen (MBN) was potential RD (PRD) > light RD (LRD) > moderate RD (MRD) > intensive RD (IRD), whereas the changing trend of other indicators was not entirely consistent with the succession of RD. The degradation trend of ISF was basically parallel to the aggravation of RD, and the strength of ISF mean values were in the order of PRD > LRD > MRD > IRD. The TOC, MBC, and MBN could be regarded as the key indicators to evaluate the soil fertility.
Abstract. Expanding of karst rocky desertification (RD) area in southwestern China has led to destructed ecosystem and local economic development lagging behind. It is important to understand the soil fertility at RD regions for the sustainable management of karst lands. The effects of the succession of RD on soil fertility were studied by investigating the stands and analyzing the soil samples with different RD grades in the central Hunan province, China, using the principal component analysis method. The results showed that the succession of RD had different impacts on soil fertility indicators. The changing trend of total organic carbon (TOC), total nitrogen (TN), available phosphorous (AP), microbial biomass carbon (MBC), and microbial biomass nitrogen (MBN) out of 19 selected indicators in different RD regions was: potential RD (PRD) > light RD (LRD) > moderate RD (MRD) > intensive RD (IRD), whereas the changing trend of other indicators was not entirely consistent with the succession of RD. The degradation trend of soil fertility was basically parallel to the aggravation of RD, and the strength of integrated soil fertility was in the order of PRD > MRD > LRD > IRD. The TOC, total phosphorus (TP), cation exchange capacity (CEC), MBC, MBN, microbial mass phosphorous (MBP), and bulk density (BD) could be regarded as the key indicators to evaluate the soil fertility due to their close correlations to the integrated fertility.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.