Exploring enhanced non-resonant di-Higgs production at the HL-LHC with neural networks
- 1Instituto Balseiro and CONICET
- 2Universidad Nacional de La Plata
Abstract
We investigate di-Higgs production in the bb̅γ γ final state at the LHC, focusing on scenarios where the gluon fusion process is enhanced by new colored scalars, which could be identified as squarks or leptoquarks. We consider two benchmarks characterized by the mass of the lightest colored scalar, BML and BMH, corresponding to 464 GeV and 621 GeV, respectively. Using Monte Carlo simulations for both the signal and the dominant backgrounds, we perform a discovery analysis with deep neural networks, exploring various architectures and input variables. Our results show that the discrimination power is maximized by employing two dedicated classifiers, one trained against QCD backgrounds and another against backgrounds involving single-Higgs processes. Furthermore, we demonstrate that including high-level features – such as the invariant masses mγ γ , mbb, and mhh, as well as the transverse momenta and angular separations of the photon and b-jet pairs – significantly improves the performance compared to using only low-level features as the invariant mass and momenta of the final particles. For the latter case, we find that architectures processing photon and b-jet variables separately can enhance the significance for BMH. Projecting for an integrated luminosity of 3 ab−1, we obtain a significance of 7.2 for BML, while it drops to 3.1 for BMH. In the particular case of BML, discovery level significance can be reached at 1.4 ab−1. We also estimate how these results could be modified by potential systematic uncertainties.
Related
- Edge-enhanced diffractive neural networks based on spin-multiplexed nonlocal metasurfaces
- Heterogeneity-enhanced stochastic resonance improves liquid-state computing in delayed spiking neural networks
- Resonant enhanced detection of the higher-order modes of a locked cavity
- Seeing in the Dark: Gravitational Waves from Resonant Production of Dark Photon Dark Matter
- Production of lepton-flavor-violating scalars through resonant positive-muon annihilation on atomic electrons
BibTeX
@article{rold2026exploring,
title = {{Exploring enhanced non-resonant di-Higgs production at the HL-LHC with neural networks}},
author = {Leandro Da Rold and Manuel Epele and Anibal D. Medina and Nicolás I. Mileo and Alejandro Szynkman},
journal = {The European Physical Journal C},
year = {2026},
month = nov,
doi = {10.1140/epjc/s10052-026-16516-z},
url = {https://link.springer.com/article/10.1140/epjc/s10052-026-16516-z},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Exploring enhanced non-resonant di-Higgs production at the HL-LHC with neural networks
AU - Leandro Da Rold
AU - Manuel Epele
AU - Anibal D. Medina
AU - Nicolás I. Mileo
AU - Alejandro Szynkman
PY - 2026
DA - 2026/11/03
JO - The European Physical Journal C
DO - 10.1140/epjc/s10052-026-16516-z
UR - https://link.springer.com/article/10.1140/epjc/s10052-026-16516-z
AB - We investigate di-Higgs production in the bb̅γ γ final state at the LHC, focusing on scenarios where the gluon fusion process is enhanced by new colored scalars, which could be identified as squarks or leptoquarks. We consider two benchmarks characterized by the mass of the lightest colored scalar, BM_L and BM_H, corresponding to 464 GeV and 621 GeV, respectively. Using Monte Carlo simulations for both the signal and the dominant backgrounds, we perform a discovery analysis with deep neural networks, exploring various architectures and input variables. Our results show that the discrimination power is maximized by employing two dedicated classifiers, one trained against QCD backgrounds and another against backgrounds involving single-Higgs processes. Furthermore, we demonstrate that including high-level features – such as the invariant masses m_γ γ , m_bb, and m_hh, as well as the transverse momenta and angular separations of the photon and b-jet pairs – significantly improves the performance compared to using only low-level features as the invariant mass and momenta of the final particles. For the latter case, we find that architectures processing photon and b-jet variables separately can enhance the significance for BM_H. Projecting for an integrated luminosity of 3 ab^−1, we obtain a significance of 7.2 for BM_L, while it drops to 3.1 for BM_H. In the particular case of BM_L, discovery level significance can be reached at 1.4 ab^−1. We also estimate how these results could be modified by potential systematic uncertainties.
SN - 1434-6052
ER -