Accelerated Randomized Benchmarking

TL;DR

This paper introduces an accelerated randomized benchmarking method that significantly reduces data requirements and improves accuracy for assessing quantum control fidelities, enabling more efficient quantum system evaluation.

Contribution

It presents a novel approach that combines prior information and advanced algorithms to enhance randomized benchmarking efficiency and accuracy in quantum systems.

Findings

Achieves several orders of magnitude better accuracy with less data.

Requires an order of magnitude less data for the same fidelity estimates.

Enables online error estimation and practical application in physical devices.

Abstract

Quantum information processing offers promising advances for a wide range of fields and applications, provided that we can efficiently assess the performance of the control applied in candidate systems. That is, we must be able to determine whether we have implemented a desired gate, and refine accordingly. Randomized benchmarking reduces the difficulty of this task by exploiting symmetries in quantum operations. Here, we bound the resources required for benchmarking and show that, with prior information, we can achieve several orders of magnitude better accuracy than in traditional approaches to benchmarking. Moreover, by building on state-of-the-art classical algorithms, we reach these accuracies with near-optimal resources. Our approach requires an order of magnitude less data to achieve the same accuracies and to provide online estimates of the errors in the reported fidelities. We also show that our approach is useful for physical devices by comparing to simulations. Our results thus enable the application of randomized benchmarking in new regimes, and dramatically reduce the experimental effort required to assess control fidelities in quantum systems. Finally, our work is based on open-source scientific libraries, and can readily be applied in systems of interest.