8–10 Jul 2026
Europe/Sarajevo timezone

Recovering LHC sensitivity to GGM in photon-deficient regions with boosted objects and machine learning

Not scheduled
20m

Speaker

Mr Debabrata Sahoo (Institute of Physics Bhubaneswar)

Description

We propose a comprehensive search strategy for gluinos in the framework of General Gauge Mediation (GGM) at the high-luminosity LHC. Conventional GMSB searches at the LHC rely on a final state with multiple hard photons originating from the decay of a bino-higgsino-mixed NLSP into a gravitino LSP. As we have shown in our previous work, this strategy loses sensitivity over a significant portion of the GGM parameter space, where the cascade chain that ends at the photon-emitting NLSP is broken either by the direct decay of the gluino into a gluon and a gravitino, or by the gravitino decays of the heavier electroweakinos $\tilde\chi_{2,3}^0$ and $\tilde\chi_1^\pm$ into $W^\pm$, $Z$, or $h$ bosons accompanied by a gravitino. To restore sensitivity in these photon-deficient regions, we capture the boosted hadronic decays of the resulting electroweak bosons and top quarks using fat jets, identified by a graph-neural-network-based multi-class classifier built on the LorentzNet architecture. We organise the analysis into nine mutually exclusive signal regions characterised by photon, fat-jet, and lepton multiplicities, and use per-region boosted decision trees as the final discriminants in a binned profile-likelihood limit calculation. At $\sqrt{s} = 14$ TeV with $500~\text{fb}^{-1}$ of integrated luminosity, the combined analysis projects a $95\%$ CL expected exclusion on the gluino mass of up to $\sim 2.83$ TeV, with complementarity among the photonic, boosted-object, and dijet-plus-$E_T^{\rm miss}$ signal regions ensuring coverage across the full $(M_1, M_3)$ plane.

Authors

Mr Debabrata Sahoo (Institute of Physics Bhubaneswar) Prof. Kirtiman Ghosh (Institute of Physics Bhubaneswar) Dr Rameswar Sahu (Department of Physics and Astrophysics, University of Delhi)

Presentation materials

There are no materials yet.