Heavy neutrino mixing prospects at hadron colliders: a machine learning study

Si-Yu Chen, Yu-Peng Jiao, Shi-Yu Wang, Qi-Shu Yan, Hong-Hao Zhang, Yongchao Zhang · 2025

Read Heavy neutrino mixing prospects at hadron colliders: a machine learning study by Si-Yu Chen, Yu-Peng Jiao, Shi-Yu Wang, Qi-Shu Yan, Hong-Hao Zhang, Yongchao Zhang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-12141.txt.

Introduction

Si-Yu Chen and co-authors use machine learning to sharpen the hunt for heavy neutrinos at large particle colliders. Their 2025 study works within a left-right model where heavy neutrinos can yield two charged leptons and two jets. The lepton pair can carry the same charge or opposite charges.

Patterns in the lepton and jet data, and links among them, feed the learning models. One boosting method sorts signal from background well in both charge cases. The team then maps how far future machines could probe. At the highest energy studied, heavy neutrinos could be found at masses well above ten trillion electronvolts. Reach for their mixing strength is also charted for each lepton mix.

Catalog record

CREATORSi-Yu Chen, Yu-Peng Jiao, Shi-Yu Wang, Qi-Shu Yan, Hong-Hao Zhang, Yongchao Zhang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE63 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-12141

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Heavy neutrino mixing prospects at hadron colliders: a machine learning study Si-Yu Chen, Yu-Peng Jiao, Shi-Yu Wang, Qi-Shu Yan, Hong-Hao Zhang, Yongchao Zhang Abstract We apply machine learning to the searches of heavy neutrino mixing in the inverse seesaw in the framework of left-right symmetric model at the high-energy hadron colliders. The Majorana nature of heavy neutrinos can induce the processes p ​ p → W R ± → ℓ α ± ​ N → ℓ α ± ​ ℓ β ∓ , ± ​ j ​ j pp\to W_{R}^{\pm}\to\ell_{\alpha}^{\pm}N\to\ell_{\alpha}^{\pm}\ell_{\beta}^{\mp,\,\pm}jj , with opposite-sign (OS) and same-sign (SS) dilepton and two jets in the final state. The distributions of the charged leptons ℓ = e , μ \ell=e,\,\mu and jets and their correlations are utilized as input for machine learning analysis. It is found that for both the OS and SS processes, XGBoost can efficiently distinguish signals from the standard model backgrounds. We estimate the sensitivities of heavy neutrino mass m N m_{N} and their mixing in the OS and SS e ​ e ee , μ ​ μ \mu\mu and e ​ μ e\mu final states at s = 14 \sqrt{s}=14 TeV, 27 TeV and 100 TeV. It turns out that the heavy neutrinos can be probed up to 17.1 TeV and 19.5 TeV in the OS and SS channels, respectively. The sine of the mixing angle of heavy neutrinos can be probed up to the maximal value of 2 / 2 \sqrt{2}/2 and 0.69 in the OS and SS channels, respectively. 1 Introduction One of the foremost objectives in particle physics is to elucidate the origin of neutrino masses and mixings. A simple and straightforward approach is to introduce the heavy right-handed neutrinos along with the associated Majorana masses M_{N} and Dirac masses M_{D} . This addition allows for the generation of a small but nonzero neutrino masses: \displaystyle m_{\nu}=-M_{D}M_{N}^{-1}M_{D}^{\sf T}\,, (1) which is known as the type-I seesaw mechanism Minkowski:1977sc ; Mohapatra:1979ia ; Yanagida:1979as ; Gell-Mann:1979vob ; Glashow:1979nm . In this framework, if the Dirac masses M_{D} is at the electroweak scale, M_{N} is required to be extremely large, roughly 10^{14} GeV. Such a high-energy scale is apparently inaccessible by any collider experiments. To facilitate the observation of new physics for neutrino mass generation at the TeV scale, some ingredients have to be added to the seesaw model, for instance some discrete symmetries in the lepton sector (see reviews in e.g. Refs. Altarelli:2010gt ; Ishimori:2010au ; King:2013eh ; Petcov:2017ggy ; Feruglio:2019ybq ; Xing:2020ijf ; Chauhan:2023faf ; Ding:2023htn ; Ding:2024ozt ), or the canonical type-I seesaw mechanism has to be extended in some way. The inverse seesaw is one of such well-motivated examples, where three extra singlet fermions S under the standard model (SM) gauge group are introduced Mohapatra:1986aw ; Mohapatra:1986bd . In the inverse seesaw framework, the tiny neutrino masses are proportional to the Majorana mass parameter \mu_{S} , i.e. \displaystyle m_{\nu}=(M_{D}M_{N}^{-1})\mu_{S}(M_{D}M_{N}^{-1})^{\sf T}\,. (2) The introduction of the extra singlet fermions and the new suppression parameter \mu_{S} not only reduce the masses of heavy neutrinos to the accessible energy scale, but also allow for the retention of large Yukawa coupling coefficients. The Majorana nature of heavy neutrinos leads to distinct signatures at the high-energy colliders that are absent in the SM, e.g. the lepton number violation (LNV) signals (see reviews in e.g. Refs. FileviezPerez:2015mlm ; Cai:2017mow ). In the case of a single heavy neutrino N in the framework of type-I seesaw mechanism, the ratio of LNV to lepton number conserving (LNC) signals is expected to be one. However, in the presence of two heavy neutrinos, the ratio can differ significantly from one, depending on the heavy neutrino masses, mixing angle, the CP violating phase as well the mixing of heavy neutrinos with the active ones BhupalDev:2015kgw ; Deppisch:2015qwa ; Anamiati:2016uxp ; Das:2017hmg ; Antusch:2017ebe ; Abada:2019bac ; Drewes:2019byd ; Godbole:2020jqw ; Arbelaez:2021chf ; Schubert:2022lcp ; Drewes:2022rsk ; Batra:2023ssq . In this paper, we focus on the LNC and LNV signals induced heavy neutrinos in the inverse seesaw in the framework of left-right symmetric model (LRSM) Pati:1974yy ; Mohapatra:1974gc ; Senjanovic:1975rk . In such a framework, the heavy neutrinos can be produced at the high-energy hadron colliders from the decay of an (off-shell) W_{R} boson, i.e. through the Keung-Senjanović (KS) process Keung:1983uu pp\to W_{R}^{\pm(\ast)}\to\ell_{\alpha}^{\pm}N\to\ell_{\alpha}^{\pm}\ell_{\beta}^{\mp,\,\pm}jj\,. (3) The dilepton pair in the final state can be either opposite-sign (OS) or same-sign (SS), with the latter one revealing the Majorana nature of heavy neutrinos and constituting one of the smoking-gun signals of LRSMs. The extensive studies of this process can be found in Refs. Langacker:1984dc ; Ho:1990dt ; Datta:1992qw ; Frere:2008ct ; Schmaltz:2010xr ; Nemevsek:2011hz ; Chen:2011hc ; Nemevsek:2011aa ; Aguilar-Saavedra:2012dga ; Das:2012ii ; Han:2012vk ; Barry:2013xxa ; BhupalDev:2013ntw ; Chen:2013foz ; Helo:2013ika ; Lee:2013htl ; Deppisch:2014qpa ; Aydemir:2014ama ; Deppisch:2014zta ; Parida:2014dla ; Vasquez:2014mxa ; Gluza:2015goa ; Ng:2015hba ; BhupalDev:2015kgw ; Helo:2015ffa ; Dev:2015kca ; Gluza:2016qqv ; Lindner:2016lpp ; Mitra:2016kov ; Anamiati:2016uxp ; Senjanovic:2016bya ; Mattelaer:2016ynf ; Ruiz:2017nip ; Das:2017hmg ; Cottin:2018kmq ; Nemevsek:2018bbt ; Frank:2018ifw ; Pascoli:2018heg ; Cottin:2019drg ; BhupalDev:2019ljp ; ThomasArun:2021rwf ; Nemevsek:2023hwx (see also Refs. Helo:2013esa ; Maiezza:2015lza ; Castillo-Felisola:2015bha ; Mandal:2017tab ; Helo:2018rll ; Li:2022cuq ). In this paper, we apply machine learning to the search for two nearly degenerate heavy neutrinos and their mixing in the minimal LRSM with inverse seesaw, via the processes in Eq. ( 3 ). Some recent papers in this direction can be found in e.g. Refs. Feng:2021eke ; Liu:2023gpt ; Critchley:2024wnp (see also Refs. CMS:2024xdq ; CMS:2024hik ). For simplicity we focus only on the charged lepton flavors \ell=e,\,\mu in the final state, and require that the heavy neutrinos be lighter than the W_{R} boson. We will consider both the OS and SS dilepton signals, and the SM backgrounds are mainly from the W^{+}W^{-}jj , W^{\pm}W^{\pm}jj , Zjj , W^{\pm}Zjj and t\bar{t}jj processes, depending on the charged lepton flavors involved and whether the signal is OS or SS process. The kinetic distributions of the charged leptons, jets and the missing transverse energy for the various signals and the corresponding backgrounds are utilized as inputs for the analysis. It is found that machine learning, such as the Extreme Gradient Boosting (XGBoost), can help distinguish significantly signals from backgrounds Chen:2016btl . We perform the simulations for three benchmark high-energy pp colliders, i.e. the 14 TeV High-Luminosity Large Hadron collider (HL-LHC) with an integrated luminosity of 3 ab -1 , the 27 TeV High-Energy LHC (HE-LHC) with the luminosity of 15 ab -1 ATLAS:2019mfr , and the future 100 TeV collider, such as the Future Circular Collider (FCC-hh) and the Super Proton-Proton Collider (SPPC), with the luminosity of 30 ab -1 FCC:2018vvp ; Tang:2015qga . It turns out that the prospects of the heavy neutrino mass m_{N} and the mixing angle at the future high-energy colliders are distinct in the OS and SS signals. In particular, the heavy neutrinos N can be probed up to 3.78 (4.22) TeV, 7.9 (8.2) TeV and 17.1 (19.5) TeV at the center-of-mass energies of \sqrt{s}=14 TeV, 27 TeV and 100 TeV in the OS (SS) channel. The sine of the heavy neutrino mixing angle can be probed in the range of [0.30 0.707], [0, 0.707] and

Read the complete text