Long-Sequence Quantum Process Tomography

It is indispensable for development of a quantum computer to further improve accuracy of elemental quantum operations. Details of errors during the operations are useful information for achieving the accuracy improvement. Quantum tomography has a potential to provide such information, but its standard protocols suffer from low reliability due to its high sensitivity against state preparation and measurement (SPAM) errors. Self-consistent quantum tomography such as GST and RSCQT overcomes the low reliability, but it requires too large costs of experiments and data-processing. Here, we propose a new tomographic method with an error amplification, named long-sequence quantum process tomography. We theoretically prove that the error amplification can suppress effects of SPAM errors in arbitrary finite dimensional system, and numerically show that its implementation costs can be practical for one- and two-qubit systems. We also explain how to use the method for characterizing leakage and crosstalk errors, and discuss its practicality beyond two-qubit systems.