A high-dimensional unbinned measurement of 24 kinematic observables using a machine learning-based analysis of data recorded by the ATLAS detector

In this thesis, a precision measurement of high momentum Z → µµ events in protonproton collision data recorded during 2015-2018 by the ATLAS detector is performed.This measurement is made using a new method, MultiFold, that leverages machine learning to produce a result that is not only corrected for detector effects, but is also high-dimensional and unbinned.This marks the first measurement of its kind based Contents List of Tables ix List of Figures xi Conventions and acronyms xxii 7.7 The chosen binning for the 24 observables in this analysis. . . . . . .8.1 The p-values quantifying the agreement between the MultiFold and IBU results and the target truth pseudodata for the 24 observables. .8.2 The p-values quantifying the agreement between the MultiFold and IBU results and the target truth pseudodata for 3 derived observables.8.3 The p-values quantifying the agreement between the MultiFolded pseudodata and the target in kinematic subregions. . . . . . . . . . . . . .x 8.4 The p-values quantifying the agreement between the MultiFolded pseudodata and the target for two-dimensional distributions. . . . . . . .183 xi xiv 8.1 The differential cross section distributions for a subset of the 24 observables obtained by unfolding the pseudodata with both the MultiFold method and IBU (1).. . . . . . . . . . . . . . . . . . . . . . . . . . .159 8.2 The differential cross section distributions for a subset of the 24 observables obtained by unfolding the pseudodata with both the MultiFold method and IBU (2). . . . . . . . . . . . . . . . . . . . . . . . . . . .161 8.3 The differential cross section distributions for a subset of the 24 observables obtained by unfolding the pseudodata with both the MultiFold method and IBU (3). . . . . . . . . . . . . . . . . . . . . . . . . . . .162 8.4 The differential cross section distributions for a subset of the 24 observables obtained by unfolding the pseudodata with both the MultiFold method and IBU (4). . . . . . . . . . . . . . . . . . . . . . . . . . . .163 8.5 The differential cross section distributions for a subset of the 24 observables obtained by unfolding the pseudodata with both the MultiFold method and IBU (5). . . . . . . . . . . . . . . . . . . . . . . . . . . .164 8.6 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the pseudodata with IBU (1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .165 8.7 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the pseudodata with IBU (2). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .166 8.8 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the pseudodata with IBU (3). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .167 xv 8.9 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the pseudodata with MultiFold (1). . . . . . . . . . . . . . . . . . . . . . . . . . . . .168 8.10 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the pseudodata with MultiFold (2). . . . . . . . . . . . . . . . . . . . . . . . . . . . .169 8.11 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the pseudodata with MultiFold (3). . . . . . . . . . . . . . . . . . . . . . . . . . . . .170 8.12 The relative total uncertainty in each bin of the unfolded pseudodata result using IBU and MultiFold. . . . . . . . . . . . . . . . . . . . . .171 8.13 The differential cross section distributions for three observables derived from the MultiFolded pseudodata result. . . . . . . . . . . . . . . . .174 8.14 The differential cross section measurements obtained from the Multi-Folded pseudodata for each of the 24 observables in a kinematic subregion (1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .176 8.15 The differential cross section measurements obtained from the Multi-Folded pseudodata for each of the 24 observables in a kinematic subregion (2). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .177 8.16 The differential cross section measurements obtained from the Multi-Folded pseudodata for each of the 24 observables in a kinematic subregion (3). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .179 8.17 The differential cross section measurements obtained from the Multi-Folded pseudodata for each of the 24 observables in a kinematic subregion (4). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .180 xvi 8.18 The differential cross section measurements obtained from the Multi-Folded pseudodata for each of the 24 observables in a kinematic subregion (5). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181 8.19 Two dimensional distributions obtained from the MultiFolded pseudodata result (1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .184 8.20 Two dimensional distributions obtained from the MultiFolded pseudodata result (2). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .185 8.21 The differential cross section distributions for a subset of the 24 observables obtained by unfolding the ATLAS Run 2 data using the MultiFold method (1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .187 8.22 The differential cross section distributions for a subset of the 24 observables obtained by unfolding the ATLAS Run 2 data using the MultiFold method (2). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .188 8.23 The differential cross section distributions for a subset of the 24 observables obtained by unfolding the ATLAS Run 2 data using the MultiFold method (3). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .189 8.24 The differential cross section distributions for a subset of the 24 observables obtained by unfolding the ATLAS Run 2 data using the MultiFold method (4). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .190 8.25 The differential cross section distributions for a subset of the 24 observables obtained by unfolding the ATLAS Run 2 data using the MultiFold method (5). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .191 8.26 The differential cross section distributions for three observables derived from the MultiFolded data result. . . . . . . . . . . . . . . . . . . . .193 xvii 8.27 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the ATLAS Run 2 data with IBU (1). . . . . . . . . . . . . . . . . . . . . . . . . . . . .194 8.28 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the ATLAS Run 2 data with IBU (2). . . . . . . . . . . . . . . . . . . . . . . . . . . . .195 8.29 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the ATLAS Run 2 data with IBU (3). . . . . . . . . . . . . . . . . . . . . . . . . . . . .196 8.30 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the ATLAS Run 2 data with MultiFold (1). . . . . . . . . . . . . . . . . . . . . . . . . .197 8.31 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the ATLAS Run 2 data with MultiFold (2). . . . . . . . . . . . . . . . . . . . . . . . . .198 8.32 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the ATLAS Run 2 data with MultiFold (3). . . . . . . . . . . . . . . . . . . . . . . . . .199 A.1 MC predicted distributions compared with data (7). . . . . . . . . . .204 A.2 MC predicted distributions compared with data (8). . . . . . . . . . .205 A.3 MC predicted distributions compared with data (9). . . . . . . . . . .206 B.1 The result at the reconstructed level of reweighting the MadGraph+Pythia8FxFx sample to have only positive weights (1). . . . . . . . . . . . .208 B.2 The result at the reconstructed level of reweighting the MadGraph+Pythia8 FxFx sample to have only positive weights (2). . . . . . . . . . . . .209 xviii B.3 The result at the truth level of reweighting the MadGraph+Pythia8 FxFx sample to have only positive weights (1). . . . . . . . . . . . .210 B.4 The result at the truth level of reweighting the MadGraph+Pythia8 FxFx sample to have only positive weights (2). . . . . . . . . . . . .211 B.5 The result of reweighting the combined MC sample to match the data for use in the pseudodata production (1). . . . . . . . . . . . . . . . .212 B.6 The result of reweighting the combined MC sample to match the data for use in the pseudodata production (2). . . . . . . . . . . . . . . . .213 C.1 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the pseudodata with IBU (4). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .215 C.2 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the pseudodata with IBU (5). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .216 C.3 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the pseudodata with IBU (6). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .217 C.4 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 observables obtained by unfolding the pseudodata with IBU (7). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .218 C.5 The relative uncertainties and uncertainty correlation matrices for a subset of the 24 obser

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.