English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

By colliding protons and examining particles emitted from the collisions, the Large Hadron Collider aims to study the interactions of quarks and gluons at the highest energy accessible in a controlled experimental way. In such collisions, W bosons or top quarks which have TeV scale momentum can be accessible. Reconstructing such boosted jets are becoming important. In particular, the ability to identify original particle that decays to quarks against normal QCD jets plays a central role in various searches at high energy scale. This is typically done by the use of a single physically motivated observable constructed from the constituents of the jet. In this contribution, multiple complementary observables are combined using boosted decision trees and neural networks to increase the ability to distinguish W bosons and top quarks from light quark jets in the ATLAS experiment. Since hadronic decay ratios ofW boson and top quark are higher than leptonic decay, these are essential probes for rare phase space, especially for heavy resonance search at the Large Hadron Collider (LHC). For instance, W and top quark jets are used in diboson resonance searches [1], tt resonance search [2], and measurement of differential cross section of high momentum top quarks [3] in ATLAS [4]. In this regime, frequently W and top quark jets have much higher momenta than their rest mass, thus these can be reconstructed by a single large radius jet. Such a large radius jet tends to have similar features to QCD jets produced by the single quark or gluon at much higher rate at LHC. In order to take into account radiation pattern of the jets, physically motivated observables such as mass and the number of core energy deposition area inside the jets have been used to effectively tag W or top quark jets in ATLAS [5, 6]. Applying machine learning techniques with several high level features of the large radius jet is a natural way to expand their performance by analogy to a multivariate b-tagging technique [7] . In this study, two kinds of machine learning techniques, Deep Neural Network (DNN) implemented by Keras [8] and Boosted Decision Tree (BDT) implemented by TMVA [9] are studied. Table 1. Summary of training and testing sample statistics for each signal and background processes. Purpose Training (Top Quark) Training (W Boson) Testing Sample Signal Background Signal Background Top Signal W Signal Background Number of jets 1× 10 1× 10 7× 10 7× 10 4× 10 3× 10 1× 10 1. Training Sample All simulation samples are made by Pythia8 [10], W bosons are collected from W ′ → WZ process, top quarks are collected from Z ′ → tt and quark and gluon jets are collected from QCD dijets events. The number of events used in this study is summarized in table 1. To avoid any bias due to difference of pT distribution, training samples are reweighted to obtain flat pT spectrum. While testing samples are reweighted to background pT spectrum. Input variable candidates for BDT and DNN are gathered from high level feature proposed by a theoretical community shown in table 2 [11]. Table 2. Summary of candidate variables. The definition of variables are discussed in Ref. [11]. Observable Variable Used For Jet mass m top,W Jet pT pT top,W Energy Correlation Ratios e3, C2, D2 top,W N-subjettiness τ1,τ2,τ3 ,τ21,τ32 top,W Center of Mass Observables Fox Wolfram (R 2 ) W Splitting Measures ZCUT , √ d12, √ d23 top,W Planar Flow P W Angularity a3 W Aplanarity A W KtDR KtDR W Qw Qw top Table 3. Top-quark tagging inputs groups for DNN as used in figure 2.

Identification of hadronically decaying W bosons and top quarks using multivariate techniques at ATLAS