Finding broken gates in quantum circuits: exploiting hybrid machine learning
- PDF / 628,772 Bytes
- 8 Pages / 439.37 x 666.142 pts Page_size
- 11 Downloads / 236 Views
Finding broken gates in quantum circuits: exploiting hybrid machine learning Margarite L. LaBorde1
· Allee C. Rogers1 · Jonathan P. Dowling1,2,3,4
Received: 29 January 2020 / Accepted: 14 June 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Current implementations of quantum logic gates can be highly faulty and introduce errors. In order to correct these errors, it is necessary to first identify the faulty gates. We demonstrate a procedure to diagnose where gate faults occur in a circuit by using a hybridized quantum-and-classical K-Nearest-Neighbors (KNN) machine-learning technique. We accomplish this task using a diagnostic circuit and selected input qubits to obtain the fidelity between a set of output states and reference states. The outcomes of the circuit can then be stored to be used for a classical KNN algorithm. We numerically demonstrate an ability to locate a faulty gate in circuits with over 30 gates and up to nine qubits with over 90% accuracy. Keywords Quantum machine learning · Quantum computing · Quantum gates · Quantum algorithms
1 Introduction Quantum computers are becoming more realizable as we approach the noisy intermediate-scale quantum (NISQ) era [12]. Tools like the IBM Q-Experience allow researchers to program and simulate quantum algorithms on a real quantum computer with a small number of qubits. These quantum computers are programmed using quantum logic gates, which act on the qubits to perform different operations; however, current implementations of these gates are prone to physical faults such as
B
Margarite L. LaBorde [email protected]
1
Hearne Institute for Theoretical Physics and Department of Physics and Astronomy, Louisiana State University, Baton Rouge, LA 70803, USA
2
National Institute of Information and Communications Technology, Tokyo 184-8795, Japan
3
NYU-ECNU Institute of Physics at NYU Shanghai, Shanghai 200062, China
4
CAS-Alibaba Quantum Computing Laboratory, USTC, Shanghai 201315, China 0123456789().: V,-vol
123
230
Page 2 of 8
M. L. LaBorde et al.
extraneous phase shifts or rotations, which introduce systematic errors into the system [7,10]. Before error correction protocols can be implemented, it is necessary to identify the gate producing the error. Here, we propose a preprocessing step to diagnose gate faults—without altering the circuit itself—by utilizing machine learning. Machine-learning techniques are powerful tools for classification and pattern recognition, and much work has been done to determine the potential advantages of quantum machine-learning algorithms [6,14,16]. We consider a hybrid quantum-classical machine learning technique that utilizes both quantum and classical algorithms. Similar hybrid schemes have been used to achieve machine-learning capabilities for NISQ devices [8,15]. Using a hybrid technique, we harness the computational advantage of quantum systems while utilizing more freely available classical resources such as memory. Here, we consider a machine-learning algorithm known as K-Near
Data Loading...
