Cell-level Transformers classify collimated ALP photon-jets versus single photons with AUC 0.98 and regress diphoton mass to ~64 MeV, beating shower-shape and other ML baselines in an ATLAS-like GEANT4 simulation.
Rays of light from the LHC
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
We consider models for the di-photon resonance observed at ATLAS (with 3.6 fb^{-1}) and CMS (with 2.6 fb^{-1}). We find there is no conflict between the signal reported at 13 TeV, and the constraints from both experiments at 8 TeV with 20.3 fb^{-1}. We make a simple argument for why adding only one new resonance to the standard model (SM) is not sufficient to explain the observation. We explore four viable options: (i): resonance production and decay through loops of messenger fermions or scalars; (ii): a resonant messenger which decays to the di-photon resonance + X; (iii): an edge configuration where A -> B gamma -> C gamma gamma, and (iv): Hidden Valley-like models where the resonance decays to a pair of very light (sub-GeV) states, each of which in turn decays to a pair of collimated photons that cannot be distinguished from a single photon. Since in each case multiple new states have been introduced, a wealth of signatures is expected to ensue at Run-2 of LHC.
fields
hep-ph 1years
2026 1verdicts
ACCEPT 1representative citing papers
citing papers explorer
-
Transformer-based machine learning using low-level calorimeter signals for collimated photon identification at collider experiments
Cell-level Transformers classify collimated ALP photon-jets versus single photons with AUC 0.98 and regress diphoton mass to ~64 MeV, beating shower-shape and other ML baselines in an ATLAS-like GEANT4 simulation.