|
|
A new method of constructing adversarial examples for quantum variational circuits |
Jinge Yan(颜金歌)1,2, Lili Yan(闫丽丽)1,2,†, and Shibin Zhang(张仕斌)1,2 |
1 School of Cybersecurity, Chengdu University of Information Technology, Sichuan 610000, China; 2 Advanced Cryptography and System Security Key Laboratory of Sichuan Province, Sichuan 610000, China |
|
|
Abstract A quantum variational circuit is a quantum machine learning model similar to a neural network. A crafted adversarial example can lead to incorrect results for the model. Using adversarial examples to train the model will greatly improve its robustness. The existing method is to use automatic differentials or finite difference to obtain a gradient and use it to construct adversarial examples. This paper proposes an innovative method for constructing adversarial examples of quantum variational circuits. In this method, the gradient can be obtained by measuring the expected value of a quantum bit respectively in a series quantum circuit. This method can be used to construct the adversarial examples for a quantum variational circuit classifier. The implementation results prove the effectiveness of the proposed method. Compared with the existing method, our method requires fewer resources and is more efficient.
|
Received: 27 June 2022
Revised: 24 September 2022
Accepted manuscript online: 19 October 2022
|
PACS:
|
03.67.-a
|
(Quantum information)
|
|
03.67.Ac
|
(Quantum algorithms, protocols, and simulations)
|
|
03.67.Lx
|
(Quantum computation architectures and implementations)
|
|
Fund: Project supported by the National Natural Science Foundation of China (Grant Nos. 62076042 and 62102049), the Natural Science Foundation of Sichuan Province (Grant No. 2022NSFSC0535), the Key Research and Development Project of Sichuan Province (Grant Nos. 2021YFSY0012 and 2021YFG0332), the Key Research and Development Project of Chengdu (Grant No. 2021-YF05-02424-GX), and the Innovation Team of Quantum Security Communication of Sichuan Province (Grant No. 17TD0009). |
Corresponding Authors:
Lili Yan
E-mail: yanlili@cuit.edu.cn
|
Cite this article:
Jinge Yan(颜金歌), Lili Yan(闫丽丽), and Shibin Zhang(张仕斌) A new method of constructing adversarial examples for quantum variational circuits 2023 Chin. Phys. B 32 070304
|
[1] Biamonte J, Wittek P, Pancotti N, Rebentrost P, Wiebe N and Lloyd S 2017 Nature 549 195 [2] Sarma D S, Deng D L and Duan L M 2019 Phys. Today 72 48 [3] Jordan M I and Mitchell T M 2015 Science 349 255 [4] LeCun Y, Bengio Y and Hinton G 2015 Nature 521 436 [5] Chang C C and Lin C J ACM Trans. Intell. Syst. Technol. 2 1 [6] Ji A B, Pang J H and Qiu H J 2010 Expert. Syst. Appl. 37 3495 [7] Knill E, Laflamme R and Milburn G J 2001 Nature. 409 46 [8] Wiebe N, Braun D and Lloyd S 2012 Phys. Rev. Lett. 109 050505 [9] Torlai G, Mazzola G, Carrasquilla J, Troyer M, Melko R and Carleo G 2018 Nat. Phys. 14 447 [10] Dallaire-Demers P L and Killoran N 2018 Phys. Rev. A 98 012324 [11] Harrow A W, Hassidim A and Lloyd S 2009 Phys. Rev. Lett. 103 150502 [12] Sun J and Lu S F 2020 Chin. Phys. B 29 100303 [13] Meng Y, Mei F, Chen G and Jia S T 2020 Chin. Phys. B 29 070501 [14] Rebentrost P, Mohseni M and Lloyd S 2014 Phys. Rev. Lett. 113 130503 [15] Li Z K, Liu X M, Xu N Y and Du J F 2015 Phys. Rev. Lett. 114 140504 [16] Havlíček V, Córcoles A D, Temme K, Harrow A W, Kandala A, Chow J M and Gambetta J M 2019 Nature 567 209 [17] Cong I and Duan L M 2016 New J. Phys. 18 073011 [18] Schmidhuber J 2015 Neural Networks 61 85 [19] Liu J, Lim K H, Wood K L, Huang W, Guo C and Huang H L 2021 Sci. China Phys. Mech. 64 290311 [20] Cong I, Choi S and Lukin M D 2019 Nat. Phys. 15 1273 [21] Shafee F 2007 Eng. Appl. Artif. Intel. 20 429 [22] Li P C and Li S Y 2008 J. Syst. Eng. Electron. 19 167 [23] Benedetti M, Lloyd E, Sack S and Fiorentini M 2019 Quantum Sci. Technol. 4 019601 [24] Yu X M, Tan X S, Yu H F and Yu Y 2018 Acta Phys. Sin 67 220302 (in Chinese) [25] Li X Q, Zhao Y F, Tang Y N and Yang W J 2018 Acta Phys. Sin 67 070302 (in Chinese) [26] Huang H L, Du Y X, Gong M, et al. 2021 Phys. Rev. Appl. 16 024051 [27] Harrigan M P, Sung K J, Neeley M, et al. 2021 Nat. Phys. 17 332 [28] Nilesh D, Pedro D M, Sumit S, et al. 2004 Proceedings of the tenth ACM SIGKDD international conference on Knowledge discovery and data mining, August 22-25, 2004, Seattle, USA, p. 99 [29] Szegedy C, Zaremba W, Sutskever I, Bruna J, Erhan D, Goodfellow I and Fergus R 2014 arXiv: 1312.6199 [30] Miyato T, Andrew M and Goodfellow I 2021 arxiv: 1605.07725 [31] GoodFellow I, Shlens J and Szegedy C 2015 arxiv: 1412.6572 [32] Liu N N and Wittek P 2020 Phys. Rev. A 101 062331 [33] Neidinger R 2010 SIAM Rev. 52 545 [34] Lu S, Duan L M and Deng D L 2020 Phys. Rev. Res. 2 033212 [35] Schuld M and Petruccione F 2018 Supervised Learning with Quantum Computers pp. 1-287 [36] Mitarai K, Negoro M, Kitagawa and Fujii K 2018 Phys. Rev. A 98 032309 [37] https://github.com/PennyLaneAI/pennylane [38] http://yann.lecun.com/exdb/mnist/ |
No Suggested Reading articles found! |
|
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
Altmetric
|
blogs
Facebook pages
Wikipedia page
Google+ users
|
Online attention
Altmetric calculates a score based on the online attention an article receives. Each coloured thread in the circle represents a different type of online attention. The number in the centre is the Altmetric score. Social media and mainstream news media are the main sources that calculate the score. Reference managers such as Mendeley are also tracked but do not contribute to the score. Older articles often score higher because they have had more time to get noticed. To account for this, Altmetric has included the context data for other articles of a similar age.
View more on Altmetrics
|
|
|