Playing Tag with ANN: Boosted Top Identification with Pattern Recognition

Abstract

Many searches for physics beyond the Standard Model at the Large Hadron Collider (LHC) rely on top tagging algorithms, which discriminate between boosted hadronic top quarks and the much more common jets initiated by light quarks and gluons. We note that the hadronic calorimeter (HCAL) effectively takes a "digital image" of each jet, with pixel intensities given by energy deposits in individual HCAL cells. Viewed in this way, top tagging becomes a canonical pattern recognition problem. With this motivation, we present a novel top tagging algorithm based on an Artificial Neural Network (ANN), one of the most popular approaches to pattern recognition. The ANN is trained on a large sample of boosted tops and light quark/gluon jets, and is then applied to independent test samples. The ANN tagger demonstrated excellent performance in a Monte Carlo study: for example, for jets with p_T in the 1100-1200 GeV range, 60% top-tag efficiency can be achieved with a 4% mis-tag rate. We discuss the physical features of the jets identified by the ANN tagger as the most important for classification, as well as correlations between the ANN tagger and some of the familiar top-tagging observables and algorithms.

20 pages, 9 figures

References in corpus (22)

• A Brief Introduction to PYTHIA 8.1
• Jet substructure as a new Higgs search channel at the LHC
• MadEvent: Automatic Event Generation with MadGraph
• Identifying Boosted Objects with N-subjettiness
• Recombination Algorithms and Jet Substructure: Pruning as a Tool for Heavy Particle Searches
• Jet Substructure at the Tevatron and LHC: New results, new tools, new benchmarks
• A Precision Measurement of the Mass of the Top Quark
• Stop Reconstruction with Tagged Tops
• Strategies to Identify Boosted Tops
• Substructure of high-p_T Jets at the LHC
• Finding top quarks with shower deconstruction
• Finding physics signals with shower deconstruction
• Automation of the matrix element reweighting method
• Search for resonances decaying into top-quark pairs using fully hadronic decays in pp collisions with ATLAS at sqrt(s) = 7 TeV
• Template Overlap Method for Massive Jets
• Jet Substructure Without Trees
• A search for ttbar resonances in lepton+jets events with highly boosted top quarks collected in pp collisions at sqrt(s) = 7 TeV with the ATLAS detector
• Boosting Top Partner Searches in Composite Higgs Models
• High Multiplicity Searches at the LHC Using Jet Masses
• Top Partner Discovery in the $T\to tZ$ channel at the LHC
• Boosted Event Topologies from TeV Scale Light Quark Composite Partners
• Boosted Tops from Gluino Decays

Cited by in corpus (9)

• Classification without labels: Learning from mixed samples in high energy physics
• Jet Substructure Classification in High-Energy Physics with Deep Neural Networks
• New Angles on Energy Correlation Functions
• Novel Jet Observables from Machine Learning
• Deep learning jet modifications in heavy-ion collisions
• Opening the black box of neural nets: case studies in stop/top discrimination
• Shared Data and Algorithms for Deep Learning in Fundamental Physics
• Mu_x boosted-bottom-jet tagging and Z-prime boson searches
• Top quark physics in the Large Hadron Collider era