Efficient and robust estimation of many-qubit Hamiltonians

Abstract

Characterizing the interactions and dynamics of quantum mechanical systems is an essential task in the development of quantum technologies. We propose an efficient protocol based on the estimation of the time derivatives of few qubit observables using polynomial interpolation for characterizing the underlying Hamiltonian dynamics and Markovian noise of a multi-qubit device. For finite range dynamics, our protocol exponentially relaxes the necessary time resolution of the measurements and quadratically reduces the overall sample complexity compared to previous approaches. Furthermore, we show that our protocol can characterize the dynamics of systems with algebraically decaying interactions. The implementation of the protocol requires only the preparation of product states and single-qubit measurements. Furthermore, we develop a shadow tomography method for quantum channels that is of independent interest. This protocol can be used to parallelize to learn the Hamiltonian, rendering it applicable for the characterization of both current and future quantum devices.

37 pages, 4 figures. Corrected gap in the proof of previous classical shadow process tomography protocols, included dissipation in the numerics, and improved presentation

References in corpus (8)

• Supplementary information for "Quantum supremacy using a programmable superconducting processor"
• Quantum computational advantage using photons
• Robust Online Hamiltonian Learning
• Suppressing qubit dephasing using real-time Hamiltonian estimation
• Hamiltonian tomography in an access-limited setting without state initialization
• Hamiltonian identifiability assisted by single-probe measurement
• Identification of open quantum systems from observable time traces
• High-accuracy Hamiltonian learning via delocalized quantum state evolutions