Estimating the fidelity of T gates using standard interleaved randomized benchmarking

TL;DR

This paper introduces a modified interleaved randomized benchmarking protocol that accurately estimates the fidelity of T gates in quantum computing, maintaining the advantages of Clifford 2-designs and efficient classical simulation.

Contribution

It presents a simple variation of RB that estimates T gate fidelity without losing the benefits of 2-designs, unlike previous methods.

Findings

Accurately estimates T gate fidelity using the modified protocol.

Retains the advantages of Clifford 2-designs and efficient classical simulation.

Provides a practical method for universal quantum computing error characterization.

Abstract

Randomized benchmarking (RB) is an important protocol for robustly characterizing the error rates of quantum gates. The technique is typically applied to the Clifford gates since they form a group that satisfies a convenient technical condition of forming a unitary 2-design, in addition to having a tight connection to fault-tolerant quantum computing and an efficient classical simulation. In order to achieve universal quantum computing one must add at least one additional gate such as the T gate (also known as the $\pi$/8 gate). Here we propose and analyze a simple variation of the standard interleaved RB protocol that can accurately estimate the average fidelity of the T gate while retaining the many advantages of a unitary 2-design and the fidelity guarantees that such a design delivers, as well as the efficient classical simulation property of the Clifford group. Our work complements prior methods that have succeeded in estimating T gate fidelities, but only by relaxing the 2-design constraint and using a more complicated data analysis.