Machine-learning discovery potential for a heavy Higgs boson in e+e-→ ZH at a future lepton collider
Abstract
A study of the discovery potential for a CP-even heavy neutral Higgs boson H is presented in the Type-I two-Higgs-doublet model (2HDM) at a future e+e- collider with √s=500 GeV, focusing on the process e+e-→ ZH with H→ WW, and with Z→ℓℓ or Z→νν. The final states are four jets plus two leptons (4j2ℓ) and four jets plus missing transverse momentum (4j2ν). Both a Boosted Decision Tree (BDT) and a Deep Neural Network (DNN) were trained on the same inputs. The BDT was found to be substantially faster in training and inference without any loss of performance, and it is therefore the classifier used in the interpretation. Three strategies are compared: a cut-based analysis, a one-dimensional BDT selection trained per mass point, and a two-dimensional (BDT × mH) template analysed with a profile likelihood. The 2D template improves the expected discovery significance by a factor of $2.5−−3.6$ in the 4j2ℓ channel and by a factor of $5−−18$ in the 4j2ν channel relative to the cut-based analysis, with the largest gains at high mH where the cut-based mass window loses efficiency. We translate the expected exclusions into 95% CL upper limits on σ× BR and discuss the implications for the Type-I 2HDM parameter space.
Related
- Fast BIB simulation at a future Muon Collider with generative machine learning
- Machine Learning for Invisible Dark Boson Searches at the Electron-Ion Collider
- Transformer-based machine learning using low-level calorimeter signals for collimated photon identification at collider experiments
- Fast prediction of electron-beam parameters using transfer maps and machine learning in a digital model of a MeV γγ collider
- Decoding complexity through machine learning is redefining scientific discovery
BibTeX
@misc{hashemi2026machinelearning,
title = {{Machine-learning discovery potential for a heavy Higgs boson in $e^+e^-\to ZH$ at a future lepton collider}},
author = {Majid Hashemi and Faezeh Nasirzadeh},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.02987},
eprint = {2611.02987},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.02987},
}
RIS
TY - UNPB
TI - Machine-learning discovery potential for a heavy Higgs boson in e^+e^-→ ZH at a future lepton collider
AU - Majid Hashemi
AU - Faezeh Nasirzadeh
PY - 2026
DA - 2026/11/02
PB - arXiv
DO - 10.48550/arxiv.2611.02987
UR - https://arxiv.org/abs/2611.02987
AB - A study of the discovery potential for a CP-even heavy neutral Higgs boson H is presented in the Type-I two-Higgs-doublet model (2HDM) at a future e^+e^- collider with √s=500 GeV, focusing on the process e^+e^-→ ZH with H→ WW, and with Z→ℓℓ or Z→νν. The final states are four jets plus two leptons (4j2ℓ) and four jets plus missing transverse momentum (4j2ν). Both a Boosted Decision Tree (BDT) and a Deep Neural Network (DNN) were trained on the same inputs. The BDT was found to be substantially faster in training and inference without any loss of performance, and it is therefore the classifier used in the interpretation. Three strategies are compared: a cut-based analysis, a one-dimensional BDT selection trained per mass point, and a two-dimensional (BDT × m_H) template analysed with a profile likelihood. The 2D template improves the expected discovery significance by a factor of $2.5−−3.6$ in the 4j2ℓ channel and by a factor of $5−−18$ in the 4j2ν channel relative to the cut-based analysis, with the largest gains at high m_H where the cut-based mass window loses efficiency. We translate the expected exclusions into 95% CL upper limits on σ× BR and discuss the implications for the Type-I 2HDM parameter space.
ER -