Stanchev Stancho G, Vitanov Nikolay V
Center for Quantum Technologies, Faculty of Physics, Sofia University, 5 James Bourchier blvd, 1164, Sofia, Bulgaria.
Sci Rep. 2024 Aug 6;14(1):18185. doi: 10.1038/s41598-024-68353-3.
We introduce a method to enhance the precision and accuracy of Quantum Process Tomography (QPT) by mitigating the errors caused by state preparation and measurement (SPAM), readout and shot noise. Instead of performing QPT solely on a single gate, we propose performing QPT on a sequence of multiple applications of the same gate. The method involves the measurement of the Pauli transfer matrix (PTM) by standard QPT of the multipass process, and then deduce the single-process PTM by two alternative approaches: an iterative approach which in theory delivers the exact result for small errors, and a linearized approach based on solving the Sylvester equation. We examine the efficiency of these two approaches through simulations on IBM Quantum using IBMQ_QASM_SIMULATOR. Compared to the Randomized Benchmarking type of methods, the proposed method delivers the entire PTM rather than a single number (fidelity). Compared to standard QPT, our method delivers PTM with much higher accuracy and precision because it greatly reduces the SPAM, readout and shot noise errors. We use the proposed method to experimentally determine the PTM and the fidelity of the CNOT gate on the quantum processor IBMQ_MANILA (Falcon r5.11L).
我们介绍一种方法，通过减轻由态制备和测量（SPAM）、读出及散粒噪声引起的误差，来提高量子过程层析成像（QPT）的精度和准确性。我们不是仅对单个门执行QPT，而是提议对同一门的多次应用序列执行QPT。该方法包括通过多程过程的标准QPT测量泡利转移矩阵（PTM），然后通过两种替代方法推导单过程PTM：一种迭代方法，理论上对于小误差能给出精确结果；另一种基于求解西尔维斯特方程的线性化方法。我们使用IBMQ_QASM_SIMULATOR在IBM量子计算机上通过模拟检验这两种方法的效率。与随机基准测试类型的方法相比，所提出的方法给出的是整个PTM而非单个数字（保真度）。与标准QPT相比，我们的方法能以更高的精度和准确性给出PTM，因为它极大地减少了SPAM、读出及散粒噪声误差。我们使用所提出的方法在量子处理器IBMQ_MANILA（Falcon r5.11L）上通过实验确定CNOT门的PTM和保真度。
