Characterization and Benchmarking of Quantum Computers using Cycle Benchmarking Techniques

Characterization of noisy, intermediate-scale quantum (NISQ) devices requires gathering detailed information about noise processes and their effects in quantum circuits. Cycle benchmarking is a protocol to estimate the process fidelity of a noisy quantum process. We pair this method with Pauli channel estimation to construct a description of noisy operations in a quantum circuit. We demonstrate cycle benchmarking and Pauli channel estimation in experiment on superconducting transmon qubit registers up to size 27. We use cycle benchmarking to test one- and two-qubit quantum operations and report their process fidelity. We use Pauli channel estimation to identify the noise channels that arise in experiment and develop noise models. We use the noise models in numerical simulation and compare to experiment. Our results show that cycle benchmarking and Pauli channel estimation can be used to estimate quantum device behavior within at least 14% of experiment and are at least 10% more accurate than noiseless baseline estimates.