Multiclass classification using quantum convolutional neural networks with hybrid quantum-classical learning

被引:31
作者
Bokhan, Denis [1 ,2 ]
Mastiukova, Alena S. [2 ,3 ]
Boev, Aleksey S. [2 ]
Trubnikov, Dmitrii N. [1 ]
Fedorov, Aleksey K. [2 ,3 ]
机构
[1] Lomonosov Moscow State Univ, Dept Chem, Phys Chem Div, Lab Mol Beams, Moscow, Russia
[2] Russian Quantum Ctr, Moscow, Russia
[3] Natl Univ Sci & Technol MISIS, Moscow, Russia
基金
俄罗斯科学基金会;
关键词
quantum learning; multinomial classification; paramaterized quantum circuit; variational circuits; amplitude encoding; POWER;
D O I
10.3389/fphy.2022.1069985
中图分类号
O4 [物理学];
学科分类号
0702 ;
摘要
Multiclass classification is of great interest for various applications, for example, it is a common task in computer vision, where one needs to categorize an image into three or more classes. Here we propose a quantum machine learning approach based on quantum convolutional neural networks for solving the multiclass classification problem. The corresponding learning procedure is implemented via TensorFlowQuantum as a hybrid quantum-classical (variational) model, where quantum output results are fed to the softmax activation function with the subsequent minimization of the cross entropy loss via optimizing the parameters of the quantum circuit. Our conceptional improvements here include a new model for a quantum perceptron and an optimized structure of the quantum circuit. We use the proposed approach to solve a 4-class classification problem for the case of the MNIST dataset using eight qubits for data encoding and four ancilla qubits; previous results have been obtained for 3-class classification problems. Our results show that accuracies of our solution are similar to classical convolutional neural networks with comparable numbers of trainable parameters. We expect that our finding provide a new step towards the use of quantum neural networks for solving relevant problems in the NISQ era and beyond.
引用
收藏
页数:9
相关论文
共 46 条
[21]   Quantum and classical machine learning for the classification of non-small-cell lung cancer patients [J].
Jain, Siddhant ;
Ziauddin, Jalal ;
Leonchyk, Paul ;
Yenkanchi, Shashibushan ;
Geraci, Joseph .
SN APPLIED SCIENCES, 2020, 2 (06)
[22]   Scalable quantum computing with qudits on a graph [J].
Kiktenko, E. O. ;
Nikolaeva, A. S. ;
Xu, Peng ;
Shlyapnikov, G., V ;
Fedorov, A. K. .
PHYSICAL REVIEW A, 2020, 101 (02)
[23]   Quantum computers [J].
Ladd, T. D. ;
Jelezko, F. ;
Laflamme, R. ;
Nakamura, Y. ;
Monroe, C. ;
O'Brien, J. L. .
NATURE, 2010, 464 (7285) :45-53
[24]   Recent advances for quantum classifiers [J].
Li, Weikang ;
Deng, Dong-Ling .
SCIENCE CHINA-PHYSICS MECHANICS & ASTRONOMY, 2022, 65 (02)
[25]  
Liu W-Q., 2021, ARXIV
[26]   Universal quantum simulators [J].
Lloyd, S .
SCIENCE, 1996, 273 (5278) :1073-1078
[27]  
Lloyd S, 2013, Arxiv, DOI arXiv:1307.0411
[28]   Quantum computational chemistry [J].
McArdle, Sam ;
Endo, Suguru ;
Aspuru-Guzik, Alan ;
Benjamin, Simon ;
Yuan, Xiao .
REVIEWS OF MODERN PHYSICS, 2020, 92 (01)
[29]   Kernel methods in Quantum Machine Learning [J].
Mengoni, Riccardo ;
Di Pierro, Alessandra .
QUANTUM MACHINE INTELLIGENCE, 2019, 1 (3-4) :65-71
[30]   Quantum circuits for general multiqubit gates -: art. no. 130502 [J].
Möttönen, M ;
Vartiainen, JJ ;
Bergholm, V ;
Salomaa, MM .
PHYSICAL REVIEW LETTERS, 2004, 93 (13) :130502-1