An iterative Bayesian variational quantum eigensolver with the von Mises-Fisher distribution

Abstract

We develop an iterative Bayesian variational quantum eigensolver in which the unknown ground state, parametrised as a real unit vector, is inferred through a von Mises--Fisher distribution sequentially updated from Hadamard-test measurements on a single ancilla qubit. Building on a companion paper that treated a single success outcome via exact moment matching, we generalise the scheme to condition on $m$ success outcomes for the likelihood function per iteration. Since exact moment matching becomes impractical for $m>1$, we exploit the large-concentration regime, in which a first-order expansion of the log-likelihood yields a closed-form von Mises--Fisher posterior with $m$ entering only through a multiplicative weight. We prove that the two convergence properties of the single-outcome scheme---non-decreasing overlap and non-decreasing concentration---are preserved provided $m$ lies within an explicitly characterised range. Based on the theory, we propose a measurement-based algorithm and illustrate the convergence on randomly generated three- and four-qubit Hamiltonian matrices, on $\mathrm{HeH}^+$, and on $\mathrm{H}_2$.

Abstract: arXiv descriptive metadata (CC0). Thank you to arXiv for use of its open access interoperability. This service was not reviewed or approved by, nor does it necessarily express or reflect the policies or opinions of, arXiv.