Analytic boosted boson discrimination

By A. Larkoski, I. Moult, D. Neill. 2015

149 citations recorded by the source corpus.

A bstractObservables which discriminate boosted topologies from massive QCD jets are of great importance for the success of the jet substructure program at the Large Hadron Collider. Such observables, while both widely and successfully used, have been studied almost exclusively with Monte Carlo simulations. In this paper we present the first all-orders factorization theorem for a two-prong discriminant based on a jet shape variable, D2, valid for both signal and background jets. Our factorization theorem simultaneously describes the production of both collinear and soft subjets, and we introduce a novel zero-bin procedure to correctly describe the transition region between these limits. By proving an all orders factorization theorem, we enable a systematically improvable description, and allow for precision comparisons between data, Monte Carlo, and first principles QCD calculations for jet substructure observables. Using our factorization theorem, we present numerical results for the discrimination of a boosted Z boson from massive QCD background jets. We compare our results with Monte Carlo predictions which allows for a detailed understanding of the extent to which these generators accurately describe the formation of two-prong QCD jets, and informs their usage in substructure analyses. Our calculation also provides considerable insight into the discrimination power and calculability of jet substructure observables in general.

Subfields: Nuclear and High Energy Physics

View Paper: https://www.semanticscholar.org/paper/1c5e33dba87a5e8114ef85a91e4583ae0b3a5286

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Nuclear and High Energy Physics

Citations in the corpus

Jet Substructure Classification in High-Energy Physics with Deep Neural Networks (2016, background): “In practice, while deep network classification can boost jet tagging performance, expert features offer powerful insight [24] into the validity of the simulation models used to train these networks.”
Factorization for groomed jet substructure beyond the next-to-leading logarithm (2016, background): “[26] in the resummation of jet observables that are sensitive to multi-prong substructure.”
Jet substructure at the Large Hadron Collider: A review of recent advances in theory and machine learning (2017, background): “[68] presented the first NLL accurate calculations of two-prong jet observables with the calculation of D (α) 2 for jets produced in e +e− collisions.”
Parton Shower Uncertainties in Jet Substructure Analyses with Deep Neural Networks (2016, result): “The uncertainties from the varying the parton shower for the jet images are thus of similar size to those associated other more common variables, such as those found in theoretical studies of the D2 tagger [63, 64] and those found by the ATLAS Collaboration in [65] in searches for boosted W-bosons.”
Looking Inside Jets (2019, methodology): “[143, 187, 188, 189, 190, 157]), noticeably ECFs and N -subjettiness, in both the direct QCD approach used in this book and in SCET.”
Jet substructure at the Large Hadron Collider (2018, background): “In addition to the comprehensive studies of grooming options [57, 74, 86, 94, 95], ATLAS and CMS both investigated the discrimination powers for a plethora of jet substructure variables, including N -subjettiness [110, 111], Qjet volatility [154], ratios of energy correlation functions C 2 [117], D ”
Les Houches 2017: Physics at TeV Colliders Standard Model Working Group Report (2017, methodology): “3 Towards Improved Performance and Robustness for ATLAS and CMS The strategies used by ATLAS and CMS, namely trimmed D2 [780, 787] and SoftDropped N2 [796] with DDT [800] are specific examples of the broader approaches to two-prong tagging discussed above, and therefore our study allows us to gain i”