skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.


Title: Explainable equivariant neural networks for particle physics: PELICAN
A<sc>bstract</sc> PELICAN is a novel permutation equivariant and Lorentz invariant or covariant aggregator network designed to overcome common limitations found in architectures applied to particle physics problems. Compared to many approaches that use non-specialized architectures that neglect underlying physics principles and require very large numbers of parameters, PELICAN employs a fundamentally symmetry group-based architecture that demonstrates benefits in terms of reduced complexity, increased interpretability, and raw performance. We present a comprehensive study of the PELICAN algorithm architecture in the context of both tagging (classification) and reconstructing (regression) Lorentz-boosted top quarks, including the difficult task of specifically identifying and measuring theW-boson inside the dense environment of the Lorentz-boosted top-quark hadronic final state. We also extend the application of PELICAN to the tasks of identifying quark-initiated vs. gluon-initiated jets, and a multi-class identification across five separate target categories of jets. When tested on the standard task of Lorentz-boosted top-quark tagging, PELICAN outperforms existing competitors with much lower model complexity and high sample efficiency. On the less common and more complex task of 4-momentum regression, PELICAN also outperforms hand-crafted, non-machine learning algorithms. We discuss the implications of symmetry-restricted architectures for the wider field of machine learning for physics.  more » « less
Award ID(s):
2013010
PAR ID:
10572157
Author(s) / Creator(s):
; ; ; ;
Publisher / Repository:
Journal of High Energy Physics
Date Published:
Journal Name:
Journal of High Energy Physics
Volume:
2024
Issue:
3
ISSN:
1029-8479
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. null (Ed.)
    A bstract In this paper, a new technique for reconstructing and identifying hadronically decaying τ + τ − pairs with a large Lorentz boost, referred to as the di- τ tagger, is developed and used for the first time in the ATLAS experiment at the Large Hadron Collider. A benchmark di- τ tagging selection is employed in the search for resonant Higgs boson pair production, where one Higgs boson decays into a boosted $$ b\overline{b} $$ b b ¯ pair and the other into a boosted τ + τ − pair, with two hadronically decaying τ -leptons in the final state. Using 139 fb − 1 of proton-proton collision data recorded at a centre-of-mass energy of 13 TeV, the efficiency of the di- τ tagger is determined and the background with quark- or gluon-initiated jets misidentified as di- τ objects is estimated. The search for a heavy, narrow, scalar resonance produced via gluon-gluon fusion and decaying into two Higgs bosons is carried out in the mass range 1–3 TeV using the same dataset. No deviations from the Standard Model predictions are observed, and 95% confidence-level exclusion limits are set on this model. 
    more » « less
  2. The measurement of the charge asymmetry for highly boosted top quark pairs decaying to a single lepton and jets is presented. The analysis is performed using 138 fb−1 of data collected in pp collisions at s√=13 TeV with the CMS detector during Run 2 of the Large Hadron Collider. The selection is optimized for top quark-antiquark pairs produced with large Lorentz boosts, resulting in non-isolated leptons and overlapping jets. The top quark charge asymmetry is measured for events with tt⎯⎯ invariant mass larger than 750 GeV and corrected for detector and acceptance effects using a binned maximum likelihood fit. The measured top quark charge asymmetry is in good agreement with the standard model prediction at next-to-next-to-leading order in perturbation theory with next-to-leading order electroweak corrections. Differential distributions for two invariant mass ranges are also presented. 
    more » « less
  3. A<sc>bstract</sc> A measurement is performed of Higgs bosons produced with high transverse momentum (pT) via vector boson or gluon fusion in proton-proton collisions. The result is based on a data set with a center-of-mass energy of 13 TeV collected in 2016–2018 with the CMS detector at the LHC and corresponds to an integrated luminosity of 138 fb−1. The decay of a high-pTHiggs boson to a boosted bottom quark-antiquark pair is selected using large-radius jets and employing jet substructure and heavy-flavor taggers based on machine learning techniques. Independent regions targeting the vector boson and gluon fusion mechanisms are defined based on the topology of two quark-initiated jets with large pseudorapidity separation. The signal strengths for both processes are extracted simultaneously by performing a maximum likelihood fit to data in the large-radius jet mass distribution. The observed signal strengths relative to the standard model expectation are$$ {4.9}_{-1.6}^{+1.9} $$ 4.9 1.6 + 1.9 and$$ {1.6}_{-1.5}^{+1.7} $$ 1.6 1.5 + 1.7 for the vector boson and gluon fusion mechanisms, respectively. A differential cross section measurement is also reported in the simplified template cross section framework. 
    more » « less
  4. A<sc>bstract</sc> A search is presented for a heavy resonance decaying into a Z boson and a Higgs (H) boson. The analysis is based on data from proton-proton collisions at a centre-of-mass energy of 13 TeV corresponding to an integrated luminosity of 138 fb−1, recorded with the CMS experiment in the years 2016–2018. Resonance masses between 1.4 and 5 TeV are considered, resulting in large transverse momenta of the Z and H bosons. Final states that result from Z boson decays to pairs of electrons, muons, or neutrinos are considered. The H boson is reconstructed as a single large-radius jet, recoiling against the Z boson. Machine-learning flavour-tagging techniques are employed to identify decays of a Lorentz-boosted H boson into pairs of charm or bottom quarks, or into four quarks via the intermediate H → WW*and ZZ*decays. The analysis targets H boson decays that were not generally included in previous searches using the H →$$ \textrm{b}\overline{\textrm{b}} $$ b b ¯ channel. Compared with previous analyses, the sensitivity for high resonance masses is improved significantly in the channel where at most one b quark is tagged. 
    more » « less
  5. Abstract A measurement of the jet mass distribution in hadronic decays of Lorentz-boosted top quarks is presented. The measurement is performed in the lepton + jets channel of top quark pair production ( $$\hbox {t}\overline{\hbox {t}}$$ t t ¯ ) events, where the lepton is an electron or muon. The products of the hadronic top quark decay are reconstructed using a single large-radius jet with transverse momentum greater than 400 $$\,\text {Ge}\hspace{-.08em}\text {V}$$ Ge V . The data were collected with the CMS detector at the LHC in proton-proton collisions and correspond to an integrated luminosity of 138 $$\,\text {fb}^{-1}$$ fb - 1 . The differential $$\hbox {t}\overline{\hbox {t}}$$ t t ¯ production cross section as a function of the jet mass is unfolded to the particle level and is used to extract the top quark mass. The jet mass scale is calibrated using the hadronic W boson decay within the large-radius jet. The uncertainties in the modelling of the final state radiation are reduced by studying angular correlations in the jet substructure. These developments lead to a significant increase in precision, and a top quark mass of $$173.06 \pm 0.84\,\text {Ge}\hspace{-.08em}\text {V} $$ 173.06 ± 0.84 Ge V . 
    more » « less