Identification of hadronically-decaying W bosons and top quarks using BDT and DNN in ATLAS

JONATHAN DAVID BOSSIO SOLA; ATLAS COLLABORATION

Ciudad Autónoma de Buenos Aires

Otro; Machine Learning Summer School 2018; 2018

Universidad Torcuato Di Tella

The performance of hadronically-decaying top-quark and W-boson taggers in pp collisions at s√ = 13 TeV recorded by the ATLAS experiment at the Large Hadron Collider is presented. A set of techniques, including some new to the data recorded in 2015 and 2016, are studied to determine a set of optimal cut-based taggers for use in physics analyses. A further extension is made to study the utility of combinations of substructure observables as a multivariate tagger using boosted decision trees and deep neural networks in comparison with taggers based on two-variable combinations. The performance of these taggers is studied with the data collected during 2015 and 2016 in tt¯, dijet and γ + jet event topologies.