Streamlining gate set tomography experiment designs using optimal design theory

Gate set tomography (GST) is a highly precise protocol for generating self-consistent estimates of the noise channels for all of a quantum processor’s logic gates. Traditional GST experiment designs achieve this precision by performing tomography on iteratively repeated sub-circuits called germs which are designed to amplify sensitivity to all of the parameters of a gate set. This iterative structure comes with significant overhead, however, contributing the high experimental cost of running GST. In this work we show how to leverage tools from the field of optimal experimental design, such as the analysis of the Fisher information, to identify and strip out redundant circuits from within the iterative structure of traditional GST experiment designs. In doing so we can construct designs requiring significantly fewer experiment resources without sacrificing the overall achievable precision of a gate set’s estimation. We conclude by remarking on the impact of this technique, when coupled with additional tools available for the streamlining of GST experiment designs, on the experimental feasibility of GST on systems beyond two-qubits.

SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525.