Southwest Quantum Information and Technology

Resource-efficient experiment designs for multi-qubit gate set tomography

Presenting Author: Corey Ostrove, Sandia National Laboratories
Contributing Author(s): Stefan Seritan, Matthew Grace, Kenneth Rudinger, Erik Nielsen, Kevin Young,Robin Blume-Kohout

Among the most powerful tools available for characterizing the performance of an entire quantum processor is gate set tomography (GST). GST provides high-precision estimates of all of the parameter values associated with a gate set, the set of all gates, state preparations and measurements available on a device. The experimental cost of performing full-fledged traditional GST can be out-of-reach on many platforms, however, especially when scaling to two or more qubits. High-quality GST experiments need not be as expensive as traditionally presented, however. We demonstrate this by introducing protocols for taking a traditional GST experiment design and producing heavily reduced experiment designs which achieve comparable performance using significantly fewer circuits. The first protocol, germ reduction (GR), significantly reduces the overall number of germ sequences that need to be evaluated. The second, fiducial-pair reduction (FPR), reduces the number of state preparation-measurement pairs which need to be considered. For FPR we will present two approaches, a structured approach which leverages the algebraic structure of each of the germs, and a simpler but unstructured approach where we sample a random fraction of the state preparation-measurement pairs. The efficacy of these heavily-reduced experiment designs is demonstrated analytically based on an analysis and comparison of the Fisher information matrices of these designs, as well as through direct numerical simu

(Session 5 : Thursday from 5:00 pm - 7:00 pm)