arXiv:1810.07997 · Physical Review Applied
Fast High-Fidelity Readout of a Single Trapped-Ion Qubit via Machine-Learning Methods
Zihan Ding, Jin‐Ming Cui, Yun‐Feng Huang, Chuan‐Feng Li, Tao Tu, Guang‐Can Guo
Abstract
The accuracy and speed of qubit readout can greatly affect the performance of quantum computers, which are held back by the lack of a more adaptive, accurate method for determining the system's quantum state. This study uses field-programmable gate arrays for machine-learning-assisted methods of single-qubit readout on a Yb${}^{+}$ ion-trap system, achieving 99.53% average fidelity within 171 $\ensuremath{\mu}$s per sample. The proposed scheme shows considerable advantages over traditional methods in fidelity, speed, and robustness, and is compatible with real-time readout and feedback control of qubit states.