Particle Transformers for Identifying Lorentz-Boosted Higgs Bosons Decaying to a Pair of W Bosons

The CMS Collaboration. CMS-PAS-JME-25-001, CERN (2025). Submitted to JHEP.

I was responsible for training the previous iteration of the boosted $\mathrm{H \to WW}$ tagger, based on ParticleNet, during my undergraduate years.

Particle transformer boosted HWW tagging

Abstract

A novel deep neural network classifier, the “particle transformer” (ParT), is introduced for the identification of highly Lorentz-boosted, multi-pronged jets in measurements and searches performed with the CMS detector at the LHC. Based on a self-attention mechanism that allows the model to weigh the importance of different particles, ParT is trained on a wide variety of topologies, notably demonstrating strong performance for the first time on jets originating from boosted Higgs boson decays to W bosons. The ParT algorithm achieves a tagging efficiency of $>50\%$ for such jets at a QCD multijet background efficiency of 1%, while maintaining decorrelation from the jet mass. This performance is calibrated in data collected by CMS from proton-proton collisions at 13 TeV center-of-mass energy, with a dataset corresponding to a total luminosity of $138~\mathrm{fb}^{-1}$, using the primary Lund jet planes of individual subjets. Data-to-simulation selection efficiency scale factors are measured to be in the 0.9-1 range, with relative uncertainties ranging between 7 and 23%.

Cite

@techreport{CMS-PAS-JME-25-001,
      collaboration = "CMS",
      title         = "{Particle transformers for identifying Lorentz-boosted
                       Higgs bosons decaying to a pair of W bosons}",
      institution   = "CERN",
      reportNumber  = "CMS-PAS-JME-25-001",
      address       = "Geneva",
      year          = "2025",
      url           = "https://cds.cern.ch/record/2958775",
}