Machine Learning Estimators for Jet Shapes Background Correction

Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21340%2F21%3A00353104" target="_blank" >RIV/68407700:21340/21:00353104 - isvavai.cz</a>
Result on the web
<a href="http://gams.fjfi.cvut.cz" target="_blank" >http://gams.fjfi.cvut.cz</a>
DOI - Digital Object Identifier
—

Result language
angličtina
Original language name
Machine Learning Estimators for Jet Shapes Background Correction
Original language description
Jet shapes and structure observables are the key point of interest in heavy-ion physics. As the dominant background of soft processes complicates the measurement of jet properties, it is necessary to perform the correction of the jet properties. Machine learning (ML) methods such as artiffcial neural networks (ANN), decision trees and random forests are commonly used for the regression tasks. Thus, the observed uncorrected jet properties can be used as the input variables for the ML models estimating the real corrected jet properties. In this paper, we explore the potential of ML algorithms for different combinations of input jet properties. Furthermore, we use a convolutional neural network (CNN) model to test whether the deep learning approaches can improve the estimation performance. Today, deep learning models are typically used for the neutrino experiments, such as the NOvA experiment [1, 2]. We aim to improve the background correction of the jet properties in comparison to the established area-based method.
Czech name
—
Czech description
—

Project
<a href="/en/project/LTT18001" target="_blank" >LTT18001: Collaboration on experiments in Fermi National Accelerator Laboratory, USA</a><br>
Continuities
P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)<br>S - Specificky vyzkum na vysokych skolach

Publication year
2021
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection
SPMS 2020/21 Stochastic and Physical Monitoring Systems, Proceedings of the international conferences
ISBN
978-80-01-06922-6
ISSN
—
e-ISSN
—
Number of pages
8
Pages from-to
53-60
Publisher name
České vysoké učení technické v Praze
Place of publication
Praha
Event location
Malá Skála
Event date
Jun 24, 2021
Type of event by nationality
WRD - Celosvětová akce
UT code for WoS article
—

Similar results(10)