Berlin 2024 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
QI: Fachverband Quanteninformation
QI 12: Poster I
QI 12.10: Poster
Dienstag, 19. März 2024, 11:00–14:30, Poster B
Adaptive and provably accurate estimation of quantum expectation values using the empirical Bernstein stopping rule — •Uğur Tepe1, Alexander Gresch1,2, and Martin Kliesch1,2 — 1Faculty of Mathematics and Natural Sciences, Heinrich Heine University Düsseldorf — 2Hamburg University of Technology, Institute for Quantum Inspired and Quantum Optimization, Germany
Quantum computing promises exponential speed-ups across several tasks. However, a practical quantum advantage over classical computing is yet to be sought. Potential candidate problems stem, among others, from quantum chemistry, in which a certifiably accurate result is paramount. This, in turn, already results in a massive measurement effort due to many required measurement settings and shot noise.
In this work, we utilize an adaptive stopping algorithm, the so-called empirical Bernstein stopping (EBS) algorithm in the context of the variational quantum eigensolver (VQE). EBS provides provably accurate estimates while exploiting the empirical variance information to reduce the measurement effort required to do so. We numerically benchmark EBS against non-adaptive alternatives such as Hoeffding’s inequality by setting up a VQE for the dissociation curve of the hydrogen molecule, resulting in a massively reduced measurement effort.
Keywords: Bernstein Inequality; Measurement Bottleneck; Quantum Chemistry; Variational Quantum Algorithm