Скачать или смотреть Ingo Roth - Shadow estimation of gate-set properties and from random sequences

https://nisqah2023.huji.ac.il/

Ingo Roth
Technology Innovation Institute, Abu Dhabi, UAE

With quantum devices for computing and simulation increasing in scale and complexity, there is a growing need for tools that obtain precise diagnostic information. Over the years, many procedures for characterizing quantum operations have been developed and put into practice. Well-known classes are process tomography, randomized benchmarking (RB) and gate-set tomography (GST), each obtaining a distinctly different amount of information. RB commonly returns a single number capturing an averaged property of the gate-set. It is inherently scalable, but provides comparably little diagnostic information. GST aims at full tomography but comes along with a substantial cost. This situation motivates protocols for characterizing quantum gate-sets that flexibly interpolate between these two extremes, balancing resource requirements and information gain. We present work that introduces new powerful protocols for verification, benchmarking and identification. The schemes exploit the strikingly simple primitive of random sequences sampling, common for RB. We ague that measurement data from random sequences can be of much broader use than mere benchmarking, giving rise to, as we rigorously prove, a range of surprisingly efficient protocols generally robust against state-preparation and measurement (SPAM) errors. We discuss several important applications and implications of the versatile new primitive. This includes a process or channel variant of the recently developed classical shadows technique [8] for estimating properties of channels and gate-sets. Furthermore, we can effectively emulate different randomized benchmarking protocols by suitable post-processing of the data and provide fine-grained characterization such as partial, compressive, full channel tomography, learning Pauli noise and crosstalk tomography.