Cross-verification of independent quantum devices

Today’s intermediate-scale quantum devices are on the brink of outperforming their classical counterparts. Yet these devices remain noisy, and with no trusted classical computers to check their results, we are left with a dire need to verify that they indeed perform as intended. The standard approach to this problem is to benchmark the performance of the individual building blocks using techniques such as quantum process tomography or randomized benchmarking and then extrapolate to the full system. In practice, however, such an extrapolation can be unreliable due to non-Markovian effects, unmodelled stray noise or temporally and spatially correlated noise. Hence, there is a need for full-system verification techniques able to capture every source of noise and imperfection.