Southwest Quantum Information and Technology

An efficient scalable benchmarking protocol for analog quantum computers

Presenting Author: Bharath Hebbe Madhusudhana, Max Planck Institute for Quantum Optics

Accurate and precise control of large quantum systems is paramount to achieve practical advantages on quantum devices. Therefore, benchmarking the hardware errors in quantum computers has drawn significant attention lately.Existing benchmarks for digital quantum computers involve averaging the global fidelity over a large set of quantum circuits and are therefore unsuitable for specific many-body control operations used in analog quantum devices. Moreover, average global fidelity is not the optimal figure-of-merit for the applications specific to analog devices, which often use local observables. Here, we develop a new figure-of-merit suitable for analog devices based on the reduced Choi matrix of the quantum operation. We develop an efficient, scalable protocol to completely characterize the reduced Choi matrix. We construct a set of multi-qubit mixed states that can be prepared with a high fidelity and scalability. We then use them as initial states to construct a local process tomography protocol that extracts the reduced Choi matrix. Further, we show how to classically compute the ideal reduced Choi matrix, in order to obtain the local process fidelity – our figure-of-merit. Focussing on the local process fidelities corresponding to a one and two qubit subsystems of an N-qubit system, we perform a numerical experiment, simulating common noise models to illustrate our protocol. Finally, we propose an implementation of our protocol using rydberg atoms in a tweezer array.

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