基于集成学习的功耗分析研究
|
摘要
针对单模型分类算法在训练样本数量较少时成功率偏低的问题,提出一种集成学习算法,并在DPA_Contest_V4数据集上进行实验。首先使用传统方法破解循环掩码,再使用SVM(support vector machine)、随机森林和k近邻(k-nearest neighbor,k NN)等分类算法进行训练和预测,最后将这些模型的结果集成。实验结果表明,集成模型优于单一模型,尤其当训练集中的能量迹数目较少时集成模型的成功率比单一模型高10%左右。
Aiming at the problem that the single model classification algorithm has a low success rate when the number of training samples is low, an ensemble learning algorithm is presented in this paper. The experiment was conducted by applying DPA_Contest_V4 dataset. First the traditional method is used to break the mask, and then SVM, RF and kNN classification algorithms are applied to train and predict. Finally, the results of these models are combined as an ensemble model. The experimental results show that the integrated model is superior to the single model, and the success rate of the ensemble model can be about 10% higher than that of the single model especially when the number of training samples is low.
Aiming at the problem that the single model classification algorithm has a low success rate when the number of training samples is low, an ensemble learning algorithm is presented in this paper. The experiment was conducted by applying DPA_Contest_V4 dataset. First the traditional method is used to break the mask, and then SVM, RF and kNN classification algorithms are applied to train and predict. Finally, the results of these models are combined as an ensemble model. The experimental results show that the integrated model is superior to the single model, and the success rate of the ensemble model can be about 10% higher than that of the single model especially when the number of training samples is low.
引文
[1]HOSPODAR G,GIERLICHS B,MULDER E D,et al.Machine learning in side-channel analysis:a first study[J].Journal of Cryptographic Engineering,2011,1(4):293-302.
[2]MARKOWITCH O,LERMAN L,BONTEMPI G.Side channel attack:an approach based on machine learning[C]//International Workshop on Constructive Side-Channel Analysis and Security Design.[S.l.]:Springer-verlag,2011:29-41.
[3]LERMAN L,POUSSIER R,BONTEMPI G,et al.Template attacks vs.machine learning revisited(and the curse of dimensionality in side-channel analysis)[C]//Constructive Side-Channel Analysis and Secure Design,International Workshop,Cosade 2015.Berlin,Germany:[s.n.],2015:20-33.
[4]周志华.机器学习[M].北京:清华大学出版社,2016.ZHOU Zhi-hua.Machine learning[M].Beijing:Tsinghua University Press,2016.
[5]BARTKEWITZ T,LEMKE-RUST K.Efficient template attacks based on probabilistic multi-class support vector machines[C]//International Conference on Smart Card Research and Advanced Applications.[S.l.]:Springer-Verlag,2012:263-276.
[6]TELECOM ParisTech SEN research group.DPA Contest(4th edition)[EB/OL].[2017-5-12].http://www.dpacontest.org/V4/.
[7]LERMAN L,MEDEIROS S F,BONTEMPI G,et al.Amachine learning approach against a masked AES[M]//Smart Card Research and Advanced Applications.[S.l.]:Springer,2013:61-75.
[8]ZENG Z,GU D,LIU J,et al.An Improved side-channel attack based on support vector machine[C]//10th International Conference on Computational Intelligence and Security.[S.l.]:IEEE,2014:676-680.
[9]邓高明,张鹏,赵强,等.基于PCA和SVM的电磁模板分析攻击[J].计算机测量与控制,2009,17(9):1837-1839.DENG Gao-ming,ZHANG Peng,ZHAO Qiang,et al.Electromagnetic template analysis with PCA and SVM[J].Computer Measurement&Control,2009,17(9):1837-1839.
[10]RECHBERGER C,OSWALD E.Practical template attacks[J].Lecture Notes in Computer Science,2005,3325:440-456.
[2]MARKOWITCH O,LERMAN L,BONTEMPI G.Side channel attack:an approach based on machine learning[C]//International Workshop on Constructive Side-Channel Analysis and Security Design.[S.l.]:Springer-verlag,2011:29-41.
[3]LERMAN L,POUSSIER R,BONTEMPI G,et al.Template attacks vs.machine learning revisited(and the curse of dimensionality in side-channel analysis)[C]//Constructive Side-Channel Analysis and Secure Design,International Workshop,Cosade 2015.Berlin,Germany:[s.n.],2015:20-33.
[4]周志华.机器学习[M].北京:清华大学出版社,2016.ZHOU Zhi-hua.Machine learning[M].Beijing:Tsinghua University Press,2016.
[5]BARTKEWITZ T,LEMKE-RUST K.Efficient template attacks based on probabilistic multi-class support vector machines[C]//International Conference on Smart Card Research and Advanced Applications.[S.l.]:Springer-Verlag,2012:263-276.
[6]TELECOM ParisTech SEN research group.DPA Contest(4th edition)[EB/OL].[2017-5-12].http://www.dpacontest.org/V4/.
[7]LERMAN L,MEDEIROS S F,BONTEMPI G,et al.Amachine learning approach against a masked AES[M]//Smart Card Research and Advanced Applications.[S.l.]:Springer,2013:61-75.
[8]ZENG Z,GU D,LIU J,et al.An Improved side-channel attack based on support vector machine[C]//10th International Conference on Computational Intelligence and Security.[S.l.]:IEEE,2014:676-680.
[9]邓高明,张鹏,赵强,等.基于PCA和SVM的电磁模板分析攻击[J].计算机测量与控制,2009,17(9):1837-1839.DENG Gao-ming,ZHANG Peng,ZHAO Qiang,et al.Electromagnetic template analysis with PCA and SVM[J].Computer Measurement&Control,2009,17(9):1837-1839.
[10]RECHBERGER C,OSWALD E.Practical template attacks[J].Lecture Notes in Computer Science,2005,3325:440-456.