Layer, Lukas (2022) Inference Aware Neural Optimization for Top Pair Cross-Section Measurements with CMS Open Data. [Tesi di dottorato]


Download (11MB) | Preview
[error in script] [error in script]
Item Type: Tesi di dottorato
Resource language: English
Title: Inference Aware Neural Optimization for Top Pair Cross-Section Measurements with CMS Open Data
Date: 8 April 2022
Number of Pages: 201
Institution: Università degli Studi di Napoli Federico II
Department: Fisica
Dottorato: Fisica
Ciclo di dottorato: 33
Coordinatore del Corso di dottorato:
Dorigo, TommasoUNSPECIFIED
Iorio, Alberto Orso MariaUNSPECIFIED
Date: 8 April 2022
Number of Pages: 201
Keywords: Top Physics, Machine Learning, Differentiable Programming
Settori scientifico-disciplinari del MIUR: Area 02 - Scienze fisiche > FIS/01 - Fisica sperimentale
Date Deposited: 13 Apr 2022 10:36
Last Modified: 07 Jun 2023 10:48

Collection description

In recent years novel inference techniques have been developed based on the construction of summary statistics with neural networks by minimizing inference-motivated losses via automatic differentiation. The inference-aware summary statistics aim to be optimal with respect to the statistical inference goal of high energy physics analysis by accounting for the effects of nuisance parameters during the model training. One such technique is INFERNO (P. de Castro and T. Dorigo, Comp.\ Phys.\ Comm.\ 244 (2019) 170) which was shown on toy problems to outperform classical summary statistics for the problem of confidence interval estimation in the presence of nuisance parameters. In this thesis the algorithm is extended to common high energy physics problems based on a differentiable interpolation technique. In order to test and benchmark the algorithm in a real-world application, a complete, systematics-dominated analysis of the CMS experiment, "Measurement of the top-quark pair production cross section in the tau+jets channel in pp collisions at sqrt(s) = 7 TeV" (CMS Collaboration, The European Physical Journal C, 2013) is reproduced with CMS Open Data. The application of the INFERNO-powered neural network architecture to this analysis demonstrates the potential to reduce the impact of systematic uncertainties in real LHC analysis.


Downloads per month over past year

Actions (login required)

View Item View Item