Pseudo-dimension of quantum circuits

被引:30
作者
Caro, Matthias C. [1 ]
Datta, Ishaun [1 ,2 ]
机构
[1] Tech Univ Munich, Dept Math, Munich, Germany
[2] Stanford Univ, Inst Computat & Math Engn, Stanford, CA USA
基金
美国国家科学基金会;
关键词
Quantum computing; Computational learning theory; Complexity theory; UNIFORM-CONVERGENCE; VC DIMENSION; LEARNABILITY; BOUNDS;
D O I
10.1007/s42484-020-00027-5
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We characterize the expressive power of quantum circuits with the pseudo-dimension, a measure of complexity for probabilistic concept classes. We prove pseudo-dimension bounds on the output probability distributions of quantum circuits; the upper bounds are polynomial in circuit depth and number of gates. Using these bounds, we exhibit a class of circuit output states out of which at least one has exponential gate complexity of state preparation, and moreover demonstrate that quantum circuits of known polynomial size and depth are PAC-learnable.
引用
收藏
页数:14
相关论文
共 26 条
[1]  
Aaronson S, 2018, ONLINE LEARNING QUAN
[2]  
Aaronson S, 2016, ELECT C COMPUTATIONA, V23, P109
[3]   The learnability of quantum states [J].
Aaronson, Scott .
PROCEEDINGS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES, 2007, 463 (2088) :3089-3114
[4]   Shadow Tomography of Quantum States [J].
Aaronson, Scott .
STOC'18: PROCEEDINGS OF THE 50TH ANNUAL ACM SIGACT SYMPOSIUM ON THEORY OF COMPUTING, 2018, :325-338
[5]   Scale-sensitive dimensions, uniform convergence, and learnability [J].
Alon, N ;
BenDavid, S ;
CesaBianchi, N ;
Haussler, D .
JOURNAL OF THE ACM, 1997, 44 (04) :615-631
[6]  
[Anonymous], 2012, Convergence of Stochastic Processes
[7]   Function learning from interpolation [J].
Anthony, M ;
Bartlett, PL .
COMBINATORICS PROBABILITY & COMPUTING, 2000, 9 (03) :213-225
[8]  
Arunachalam Srinivasan, 2017, ACM SIGACT News, V48, P41, DOI 10.1145/3106700.3106710
[9]  
Bartlett P. L., 2003, Journal of Machine Learning Research, V3, P463, DOI 10.1162/153244303321897690
[10]   Prediction, learning, uniform convergence, and scale-sensitive dimensions [J].
Bartlett, PL ;
Long, PM .
JOURNAL OF COMPUTER AND SYSTEM SCIENCES, 1998, 56 (02) :174-190