Sample-efficient estimation of entanglement entropy through supervised learning

• Verfasst von: Rieger, Maximilian [VerfasserIn]
• Verfasst von: Reh, Moritz [VerfasserIn]
• Verfasst von: Gärttner, Martin [VerfasserIn]
• Titel: Sample-efficient estimation of entanglement entropy through supervised learning
• Verf.angabe: Maximilian Rieger, Moritz Reh, and Martin Gärttner
• E-Jahr: 2024
• Jahr: January 2024
• Umfang: 6 S.
• Illustrationen: Illustrationen
• Fussnoten: Veröffentlicht: 2. Januar 2024 ; Gesehen am 12.07.2024
• Titel Quelle: Enthalten in: Physical review
• Ort Quelle: Woodbury, NY : Inst., 2016
• Jahr Quelle: 2024
• Band/Heft Quelle: 109(2024), 1 vom: Jan., Artikel-ID 012403, Seite 1-6
• ISSN Quelle: 2469-9934
• DOI: doi:10.1103/PhysRevA.109.012403
• Datenträger: Online-Ressource
• Sprache: eng
• K10plus-PPN: 1895227267

Abstract

We explore a supervised machine-learning approach to estimate the entanglement entropy of multiqubit systems from few experimental samples. We put a particular focus on estimating both aleatoric and epistemic uncertainty of the network's estimate and benchmark against the best-known conventional estimation algorithms. For states that are contained in the training distribution, we observe convergence in a regime of sample sizes in which the baseline method fails to give correct estimates, while extrapolation only seems possible for regions close to the training regime. As a further application of our method, highly relevant for quantum simulation experiments, we estimate the quantum mutual information for nonunitary evolution by training our model on different noise strengths.