Microscopic parametrizations for gate set tomography under coloured noise

TL;DR

This paper introduces a microscopic parametrization for gate set tomography that incorporates time-correlated noise, reducing resource requirements and improving efficiency in characterizing quantum gates affected by non-Markovian noise.

Contribution

It develops a new parametrization method based on filter functions that accounts for finite correlation times, enhancing the efficiency of gate set tomography under realistic noise conditions.

Findings

Reduced resource requirements compared to standard GST.

Improved accuracy in gate characterization under non-Markovian noise.

Enhanced sampling efficiency demonstrated through examples.

Abstract

Gate set tomography (GST) allows for a self-consistent characterization of noisy quantum information processors. The standard device-agnostic approach treats the QIPs as black boxes that are only constrained by the laws of physics, attaining full generality at a considerable resource cost: numerous circuits built from the gate set must be run in order to amplify each of the gate set parameters. In this work, we show that a microscopic parametrization of quantum gates under time-correlated noise on the driving phase, motivated by recent experiments with trapped-ion gates, reduces the required resources enabling a more efficient version of GST. By making use of the formalism of filter functions over the noise spectral densities, we discuss the minimal parametrizations of the gate set that include the effect of finite correlation times and non-Markovian quantum evolutions during the individual gates. We compare the estimated gate sets obtained by our method and the standard long-sequence GST, discussing their accuracies in terms of established metrics, as well as showcasing the advantages of the parametrized approach in terms of the sampling complexity for specific examples.