Heavy Neutrinos across the Electroweak-to-Multi-TeV Frontier via Novel ML-Enhanced Probes

TL;DR

This work targets heavy neutrinos predicted by the seesaw mechanism across a broad mass range at the HL-LHC by analyzing the inclusive $pp \to \ell \nu jj$ final state, incorporating both $s$-channel and vector-boson fusion production. It leverages gradient-boosted decision trees to distinguish signal from Standard Model backgrounds using 31 kinematic features, with separate models per heavy-neutrino mass and lepton flavor, and performs a shape-based likelihood limit using $\mathcal{L}_{\mathrm{int}} = 3000~\mathrm{fb^{-1}}$. The projected sensitivities on the mixing parameter $|V_{\ell N}|^2$ span from $O(10^{-5})$ to $O(1)$ for $m_N$ between $50~\mathrm{GeV}$ and $10~\mathrm{TeV}$, with tau-inclusive channels enhancing reach in the third generation and the potential to test lepton universality by combining flavors. Overall, the method provides a complementary, ML-enhanced pathway to probe heavy neutrinos across the electroweak-to-multi-TeV frontier, extending current limits and informing neutrino-mass-generation scenarios.

Abstract

We propose a new strategy to probe heavy neutrinos with non-universal fermion couplings at the Large Hadron Collider (LHC) using a novel production mechanism and machine-learning algorithms. Focusing on proton--proton collisions at $\sqrt{s} = 13.6~\mathrm{TeV}$, we investigate final states containing a charged lepton, missing transverse energy, and two jets. For heavy neutrino masses below $\mathcal{O}(1~\mathrm{TeV})$, production is dominated by the $s$ channel process. At higher masses, vector boson fusion becomes the dominant production mechanism, with cross sections that decrease slowly as the heavy neutrino mass increases. We simulate both signal and Standard Model background events and employ gradient-boosted decision trees to optimize event classification. Assuming an integrated luminosity of $3000~\mathrm{fb^{-1}}$, expected for the high-luminosity, and considering realistic statistical and systematic uncertainties, we find that heavy neutrinos in the mass range $50~\mathrm{GeV}$--$10~\mathrm{TeV}$ can be probed with sensitivity to the mixing parameter $|V_{\ell N}|^2$ spanning from $\mathcal{O}(10^{-5})$ to 1. This approach enhances the discovery potential for heavy neutrinos and provides a complementary pathway to existing search strategies.

Figures (12)

• Figure 1: Representative Feynman diagrams illustrating the production of $\mu \nu_\mu jj$ via $s$ channel (left) and VBF processes (right), mediated by a heavy neutrino $N$. The parameter $V_{\mu N}$ denotes the corresponding mixing parameter.
• Figure 2: Production cross sections (with $V_{\ell N} = 1$) as functions of the heavy neutrino mass $m_N$ at a center-of-mass energy of $\sqrt{s} = 13.6~\mathrm{TeV}$. The red curve shows the cross section for the process $pp \to \mu N jj$. The blue curve corresponds to the inclusive cross section for $pp \to \mu \nu_\mu jj$ while the green curve denotes the VBF-induced cross section for the same-sign dilepton channel $pp \to \mu \mu jj$, as used in Refs. CMS:2022hvhATLAS:2024rzi. The three curves converge near $m_N \simeq 1~\mathrm{TeV}$, beyond which the VBF contribution becomes dominant. Our numerical results are consistent with Ref. Fuks:2020att.
• Figure 3: BDT output distributions (normalized to 1) for signal and background events at heavy neutrino masses of 0.1 TeV (left), 1 TeV (middle), and 10 TeV (right) for $\ell = \mu$. The vertical axis shows the normalized event yield in arbitrary units, and the horizontal axis represents the BDT output score.
• Figure 4: Relative feature importance for the BDT classifier at a heavy neutrino mass of 0.1 TeV (left), 1 TeV (middle), and 10 TeV (right) for $\ell = \mu$.
• Figure 5: Distribution (normalized to 1) of the variable $M_\mathrm{T}(\ell E^\mathrm{miss}_\mathrm{T})$ for background (BKG) and signal (SIG) events corresponding to heavy neutrino masses of 0.1, 1, and 10 TeV, shown together in a single panel for $\ell = \mu$.