Skip navigation
Please use this identifier to cite or link to this item: http://arks.princeton.edu/ark:/88435/dsp01hm50tv83v
Full metadata record
DC FieldValueLanguage
dc.contributor.advisorZhandry, Mark-
dc.contributor.advisorMcConnell, Mark-
dc.contributor.advisorRabitz, Herschel-
dc.contributor.authorLee, Juneseo-
dc.date.accessioned2021-07-26T13:03:15Z-
dc.date.available2021-07-26T13:03:15Z-
dc.date.created2021-04-26-
dc.date.issued2021-07-26-
dc.identifier.urihttp://arks.princeton.edu/ark:/88435/dsp01hm50tv83v-
dc.description.abstractIn this work we design a class of Ansätze to solve MaxCut on a parameterized quantum circuit (PQC). Gaining inspiration from properties of quantum optimal control landscapes, we consider the presence of optimization traps as a measure of complexity for hybrid variational quantum algorithms. In particular, we analytically show that no simple Ansatz, satisfying certain criteria, can provide a superpolynomial quantum advantage in solving MaxCut while nevertheless creating entanglement. Furthermore, in order to characterize properties of Ansätze that could provide a quantum advantage, we study the role of noncommutativity in PQCs through a series of numerical experiments. Finally, we compare this notion to similar properties in classical neural networks such as nonlinearity, based on the perspective of the recent moniker for PQCs as quantum neural networks.en_US
dc.format.mimetypeapplication/pdf
dc.language.isoenen_US
dc.titleOn Assessing the Quantum Advantage for MaxCut Provided by Quantum Neural Network Ansätzeen_US
dc.typePrinceton University Senior Theses
pu.date.classyear2021en_US
pu.departmentMathematicsen_US
pu.pdf.coverpageSeniorThesisCoverPage
pu.contributor.authorid920191696
pu.certificateCenter for Statistics and Machine Learningen_US
pu.mudd.walkinNoen_US
Appears in Collections:Mathematics, 1934-2023

Files in This Item:
File Description SizeFormat 
LEE-JUNESEO-THESIS.pdf1.24 MBAdobe PDF    Request a copy


Items in Dataspace are protected by copyright, with all rights reserved, unless otherwise indicated.