From Qubits to Couplings: A Hybrid Quantum Machine Learning Framework for LHC Physics

TL;DR

This work targets measuring the Higgs self-coupling via non-resonant Higgs pair production in the HH -> bbγγ channel at 13.6 TeV. It introduces HyQML, a hybrid quantum-classical framework that uses a meta-parameter mapping network to condition a parameterized quantum circuit on event-level features, with amplitude embedding to access a high-dimensional quantum feature space. The approach yields substantial gains over both purely quantum and classical baselines, including nearly a twofold increase in discovery significance and tighter 95% CL limits on the HH cross-section, while showing robustness to substantial background uncertainties and limited statistics. The study also demonstrates that one-dimensional and two-dimensional likelihood scans of kappa_lambda and kappa_2V remain consistent with SM expectations and improve constraints compared to the baselines, highlighting the practical potential of quantum-enhanced learning for precision Higgs-sector measurements at the LHC and future colliders.

Abstract

In this paper, we propose a new Hybrid Quantum Machine Learning (HyQML) framework to improve the sensitivity of double Higgs boson searches in the $HH \to b\bar{b}γγ$ final state at $\sqrt{s}$ = 13.6 TeV. The proposed model combines parameterized quantum circuits with a classical neural network meta-model, enabling event-level features to be embedded in a quantum feature space while maintaining the optimization stability of classical learning. The hybrid model outperforms both a state-of-the-art XGBoost model and a purely quantum implementation by a factor of two, achieving an expected 95% CL upper limit on the non-resonant double Higgs boson production cross-section of $1.9\timesσ_{\text{SM}}$ and $2.1\timesσ_{\text{SM}}$ under background normalization uncertainties of 10% and 50%, respectively. In addition, expected constraints on the Higgs boson self-coupling $κ_λ$ and quartic vector-boson-Higgs coupling $κ_{2V}$ are found to be improved compared to the classical and purely quantum models.

Figures (9)

• Figure 1: The leading-order Feynman diagrams for (a–b) gluon–gluon fusion and (c–e) vector boson fusion Higgs boson pair production.
• Figure 2: Variation of (a) ggF and (b) VBF double Higgs boson cross-section as a function of $\kappa_{\lambda}$ and $\kappa_{2V}$. If one parameter varies, the remaining parameters are fixed to 1.
• Figure 3: A schematic view of the hybrid quantum machine learning model.
• Figure 4: Weighted ROC curve for the HyQML classifier for different $\lambda$ values. The ROC curve is computed using event weights on the test dataset.
• Figure 5: Two-dimensional projection of the PQC expectation values for the first and the last epochs for $\lambda$ = 0.1.