Abstract
An important aspect of the study of Quark-Gluon Plasma (QGP) in ultrarelativistic collisions of heavy ions is the ability to identify, in experimental data, a subset of the jets that were strongly modified by the interaction with the QGP. In this work, we propose studying Deep Learning techniques for this purpose. Samples of Z+jet events were simulated in vacuum (pp collisions) and medium (PbPb collisions) and used to train Deep Neural Networks with the objective of discriminating between medium- and vacuum-like jets within the medium (PbPb) sample. Dedicated Convolutional Neural Networks, Dense Neural Networks and Recurrent Neural Networks were developed and trained, and their performance was studied. Our results show the potential of these techniques for the identification of jet quenching effects induced by the presence of the QGP.
Highlights
Because quenching is a small effect overall, but rather because jets that were significantly modified are diluted within a sample dominated by those with little modification
We observe that the area under the ROC curve (AUC) obtained with the Dense Neural Network (DNN), Recurrent Neural Network (RNN) and convolutional neural network (CNN) with unnormalised images decrease around 10% for jets with pT >125 GeV, where the pT spectra are identical between the medium and vacuum categories
The outputs of the different Deep Learning (DL) architectures are nearly uncorrelated with xjZ for vacuum, which is a desired property for the tagger since events for which xjZ differs from 1 in the vacuum result from spurious effects, independent of jet quenching through interaction with the Quark-Gluon Plasma (QGP)
Summary
To understand if Deep Learning can be applied to identify jet quenching effects we will use JEWEL v2.2.0 [19], a Monte Carlo event generator that accounts for medium-induced effects during the QCD parton shower evolution. The recoiling jet will experience several scattering processes, inducing extra radiation that is emitted at finite angles While part of this radiation stays inside the jet under the form of softer fragments, collisional energy loss contributes further to the depletion of the reconstructed transverse momentum pT,jet (and xjZ) and effectively reduces the number of particles nconst since they are transported up to large radial distances in (η, φ) [29]. The criterion on the minimum jet transverse momentum induces a selection bias on the medium sample: pairs whose recoiling jet is below the cut-off will not be included These configurations are usually dominated by jets with a larger number of constituents and a wider fragmentation pattern. Since this is a powerful discriminant in itself, we preserve it as a physical benchmark against which to compare the different DL outputs
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
