Berlin 2024 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
QI: Fachverband Quanteninformation
QI 12: Poster I
QI 12.14: Poster
Tuesday, March 19, 2024, 11:00–14:30, Poster B
Fast Hamiltonian Learning using the Bayesian parameter shift rule — •Jonathan Schluck1, Lennart Bittel2, and Martin Kliesch3 — 1Heinrich-Heine University, Duesseldorf, Germany — 2Freie Universität, Berlin, Germany — 3Hamburg University of Technology, Institute for Quantum Inspired and Quantum Optimization, Germany
Hamiltonian learning commonly refers to determining unknown parameters of a Hamiltonian from measurement data. Andi Gu et al. have proposed an efficient protocol for the estimation of its coefficients in the Pauli basis. The measurements are given by time derivatives of certain expectation values. Such estimation problems can be solved using the parameter shift rule known in the context of variational quantum algorithms. In this work, we use an extension of the parameter shift rule based on Bayesian estimation to estimate the time derivatives from fewer measurements. We demonstrate numerically that this leads to a reduction of the total measurement effort of the Hamiltonian learning protocol.
Keywords: Hamiltonian learning; Quantum Heisenberg model; Bayesian parameter shift rule; Fourier series; Quantum measurement simulation