pith. sign in

arxiv: 2601.13640 · v2 · submitted 2026-01-20 · ✦ hep-ph

Enhanced sensitivity to the H to Zγ to ell^+ell^-γ decay at the LHC using machine learning and novel kinematic observables

Manisha Kumari , Amal Sarkar This is my paper

Pith reviewed 2026-05-16 13:03 UTC · model grok-4.3

classification ✦ hep-ph
keywords Higgs bosonZ gamma decaymachine learningkinematic observablesDrell-Yan backgroundLHCboosted decision treesignal background discrimination
0
0 comments X

The pith

Physics-motivated observables from the Higgs momentum and Z-gamma angle plane, combined with XGBoost, improve signal-to-background discrimination for the H to Z gamma decay.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper seeks to better isolate the rare Higgs boson decay into a Z boson and a photon, observed through lepton pairs plus a photon, from the large Drell-Yan background at the LHC. It introduces new correlated observables derived from the two-dimensional distribution of Higgs transverse momentum and the angle between the Z and the photon. These are fed into a boosted decision tree classifier, yielding higher area under the ROC curve and signal-to-background ratios reaching 2.1 percent in the electron channel and 3.4 percent in the muon channel near the Higgs mass. This approach maintains high signal efficiency while enhancing background rejection through interpretable multivariate methods.

Core claim

By deriving additional observables that encode angular and momentum differences from the (P_Higgs, θ_Zγ) plane and incorporating them into an XGBoost classifier, the analysis achieves improved performance in distinguishing the H→Zγ→ℓ⁺ℓ⁻γ signal from the Z/γ* → ℓ⁺ℓ⁻ background, with optimised rejection leading to signal-to-background ratios of 2.1% and 3.4% for electrons and muons respectively.

What carries the argument

The (P_Higgs, θ_Zγ) plane from which physics-motivated correlated observables are extracted to capture kinematic differences between signal and background processes.

If this is right

  • Enhanced discrimination allows for higher sensitivity in measurements of this rare Higgs decay mode.
  • The method maintains high signal efficiency while improving background suppression.
  • Optimised background rejection increases the signal-to-background ratio near the Higgs mass peak.
  • The technique is flexible and applicable to other rare Higgs decays and resonant searches.
  • Combining kinematic correlations with multivariate analysis leads to robust improvements in performance.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Applying this to the full LHC dataset could tighten constraints on Higgs boson properties.
  • These observables might reveal subtle differences in other beyond-Standard-Model scenarios involving similar final states.
  • Integration with other machine learning architectures could yield additional gains in classification accuracy.
  • Validation on real data would confirm if the simulated improvements translate directly to experimental sensitivity.

Load-bearing premise

The Monte Carlo simulations at 13 TeV accurately reproduce the kinematic distributions and correlations of both the signal and Drell-Yan background as they appear in actual LHC collision data.

What would settle it

Observation of significant mismodeling or discrepancies between simulated and real data distributions in the (P_Higgs, θ_Zγ) plane or the derived observables would undermine the claimed performance improvements.

read the original abstract

At LHC energies, the Drell--Yan ($Z/\gamma^{*}$) processes have a substantially large cross section. Their di-lepton ($\ell^+\ell^-$) final state contributes significantly to many resonant signal regions, making them one of the dominant backgrounds in numerous physics analyses. The study focuses on improving the discrimination and suppression of the $Z/\gamma^{*} \rightarrow \ell^{+}\ell^{-}$ background from the $H \rightarrow Z\gamma \rightarrow \ell^{+}\ell^{-}\gamma$ signal at $\sqrt{s}=13~\text{TeV}$ by leveraging Monte Carlo simulated data. The analysis introduces physics-motivated correlated observables derived from the two-dimensional $(P_{\mathrm{Higgs}}, \theta_{Z\gamma})$ plane. These observables encode differences in angular and momentum information to enhance signal--background separation while maintaining high signal efficiency. We present a multivariate analysis (MVA) employing a Boosted Decision Tree (XGBoost) classifier. By incorporating additional physics-motivated correlated observables, the classifier achieves measurable improvements in performance. A significant increase in the area under the ROC curve (AUC) is observed in both the electron and muon channels, demonstrating the effectiveness of the expanded feature set. Further, optimised background rejection using $(P_{\mathrm{Higgs}}, \theta_{Z\gamma})$ plane increases the signal-to-background ratio to 2.1\% and 3.4\% for the electron and muon channel respectively near the Higgs mass. This work demonstrates that combining kinematic correlations with interpretable multivariate techniques leads to improved sensitivity and robust background rejection. The approach is flexible and can be readily applied to a wide range of analyses, including rare Higgs decays, resonant searches, and studies beyond the Standard Model.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The paper claims that novel kinematic observables derived from the two-dimensional (P_Higgs, θ_Zγ) plane, when combined with an XGBoost classifier trained on Monte Carlo samples at √s=13 TeV, improve discrimination between the H→Zγ→ℓ⁺ℓ⁻γ signal and the Drell-Yan background, yielding higher AUC values in both electron and muon channels and optimized signal-to-background ratios of 2.1% (electrons) and 3.4% (muons) near the Higgs mass.

Significance. If the Monte Carlo modeling of angular and momentum correlations holds, the approach offers an interpretable way to enhance sensitivity to the rare H→Zγ decay while preserving signal efficiency, with potential extension to other resonant searches; the use of physics-motivated observables rather than black-box features is a positive aspect.

major comments (1)
  1. [Abstract] Abstract and results: the reported AUC gains and S/B improvements (2.1% electron, 3.4% muon) are obtained exclusively from Monte Carlo simulations; no control-region comparisons, sideband tests, or data-MC agreement studies are described to establish that the kinematic distributions and correlations in the (P_Higgs, θ_Zγ) plane are faithfully reproduced for both signal and background, which is load-bearing for any claim of enhanced LHC sensitivity.
minor comments (2)
  1. [Abstract] Abstract: the derivation of the physics-motivated observables from the (P_Higgs, θ_Zγ) plane is not detailed, nor is the cross-validation strategy or hyperparameter tuning for the XGBoost classifier.
  2. [Abstract] Abstract: no discussion of systematic uncertainties (e.g., from MC modeling or classifier training) or the statistical significance of the quoted AUC and S/B improvements is provided.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading and constructive feedback. We address the major comment below and clarify the scope of our Monte Carlo-based study.

read point-by-point responses

  1. Referee: [Abstract] Abstract and results: the reported AUC gains and S/B improvements (2.1% electron, 3.4% muon) are obtained exclusively from Monte Carlo simulations; no control-region comparisons, sideband tests, or data-MC agreement studies are described to establish that the kinematic distributions and correlations in the (P_Higgs, θ_Zγ) plane are faithfully reproduced for both signal and background, which is load-bearing for any claim of enhanced LHC sensitivity.

    Authors: We agree that the reported improvements in AUC and S/B ratios are derived solely from Monte Carlo simulations of signal and background processes at √s=13 TeV. The manuscript is a methodological study that introduces correlated observables from the (P_Higgs, θ_Zγ) plane and evaluates their utility within an XGBoost classifier using simulated samples; it does not constitute a full experimental analysis on LHC data. This simulation-only approach is standard when proposing new observables and demonstrating their discrimination power before deployment in data analyses. We will revise the abstract, introduction, and conclusions to explicitly state that the results are Monte Carlo based, to emphasize the proof-of-concept nature of the work, and to note that any claim of enhanced LHC sensitivity assumes faithful modeling of the relevant kinematic correlations (as already implied by the referee's own summary). We will also add a brief discussion highlighting the need for future data-MC validation studies in control regions. These changes clarify the scope without altering the technical results. revision: yes

Circularity Check

0 steps flagged

No circularity: standard supervised ML evaluation on independent MC test samples

full rationale

The paper defines physics-motivated observables from the (P_Higgs, θ_Zγ) plane, trains an XGBoost classifier on Monte Carlo signal and Drell-Yan background samples, and reports AUC and optimized S/B on held-out simulated events. This is the conventional non-circular workflow for assessing discrimination power; the reported gains are measured quantities, not algebraic identities or self-referential fits. No self-citations, self-definitional steps, or renamings of known results appear in the provided text. The derivation chain is self-contained against external benchmarks (MC truth labels).

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim depends on the fidelity of Monte Carlo modeling for signal and background kinematics and the assumption that the new observables provide genuine additional separation power.

axioms (1)
  • domain assumption Monte Carlo event generators accurately model the kinematic distributions and correlations for both H→Zγ signal and Drell-Yan background at 13 TeV.
    All training, testing, and performance claims rely on simulated samples without reported data validation.

pith-pipeline@v0.9.0 · 5627 in / 1480 out tokens · 66799 ms · 2026-05-16T13:03:51.428878+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

40 extracted references · 40 canonical work pages · 6 internal anchors

  1. [1]

    Glashow,Partial-symmetries of weak interactions,Nuclear Physics22(1961) 579

    S.L. Glashow,Partial-symmetries of weak interactions,Nuclear Physics22(1961) 579

    work page 1961

  2. [2]

    Weinberg,A model of leptons,Phys

    S. Weinberg,A model of leptons,Phys. Rev. Lett.19(1967) 1264

    work page 1967

  3. [3]

    Salam,Weak and Electromagnetic Interactions,Conf

    A. Salam,Weak and Electromagnetic Interactions,Conf. Proc. C680519(1968) 367

    work page 1968

  4. [4]

    Peskin and Daniel V

    M.E. Peskin and D.V. Schroeder,An Introduction to Quantum Field Theory, Addison-Wesley (1995), 10.1201/9780429503559

    work page doi:10.1201/9780429503559 1995

  5. [5]

    Donoghue, E

    J.F. Donoghue, E. Golowich and B.R. Holstein,Dynamics of the Standard Model, Cambridge University Press (1992), 10.1017/CBO9780511524370

    work page doi:10.1017/cbo9780511524370 1992

  6. [6]

    The Standard Model of Electroweak Interactions

    A. Pich,The Standard Model of Electroweak Interactions, in2010 European School of High Energy Physics, pp. 1–50, 1, 2012 [1201.0537]

    work page internal anchor Pith review Pith/arXiv arXiv 2012

  7. [7]

    Higgs,Broken symmetries and the masses of gauge bosons,Phys

    P.W. Higgs,Broken symmetries and the masses of gauge bosons,Phys. Rev. Lett.13(1964) 508

    work page 1964

  8. [8]

    Englert and R

    F. Englert and R. Brout,Broken symmetry and the mass of gauge vector mesons,Phys. Rev. Lett.13(1964) 321

    work page 1964

  9. [9]

    Higgs,Spontaneous symmetry breakdown without massless bosons,Phys

    P.W. Higgs,Spontaneous symmetry breakdown without massless bosons,Phys. Rev.145 (1966) 1156. [10]ATLAScollaboration,Observation of a new particle in the search for the standard model higgs boson with the atlas detector at the lhc,Phys. Lett. B716(2012) 1. [11]CMScollaboration,Observation of a new boson at a mass of 125 gev with the cms experiment at the l...

    work page 1966

  10. [10]

    Santos et al.,Combined measurement of differential and total cross sections in theh→γγ and theh→zz ∗ →4ℓdecay channels at √s= 13tev with the atlas detector,Phys

    S. Santos et al.,Combined measurement of differential and total cross sections in theh→γγ and theh→zz ∗ →4ℓdecay channels at √s= 13tev with the atlas detector,Phys. Lett. B 786 (2018) 114786(2018) 114. [14]ATLAS Collaborationcollaboration,Combined measurements of higgs boson production and decay using up to80 fb −1 of proton-proton collision data at√s= 13...

    work page 2018

  11. [11]

    Tumasyan et al.,A portrait of the higgs boson by the cms experiment ten years after the discovery,Nature607(2022) 60

    A. Tumasyan et al.,A portrait of the higgs boson by the cms experiment ten years after the discovery,Nature607(2022) 60

    work page 2022

  12. [12]

    Englert, A

    C. Englert, A. Freitas, M.M. Mühlleitner, T. Plehn, M. Rauch, M. Spira et al.,Precision measurements of higgs couplings: implications for new physics scales,Journal of Physics G: Nuclear and Particle Physics41(2014) 113001

    work page 2014

  13. [13]

    Englert, R

    C. Englert, R. Kogler, H. Schulz and M. Spannowsky,Higgs coupling measurements at the lhc,The European Physical Journal C76(2016) 393

    work page 2016

  14. [14]

    Degrassi and F

    G. Degrassi and F. Maltoni,Two-loop electroweak corrections to the higgs-boson decay h→γγ,Nuclear Physics B724(2005) 183

    work page 2005

  15. [15]

    Djouadi,The anatomy of electroweak symmetry breaking tome ii: The higgs bosons in the minimal supersymmetric model,Physics Reports459(2008) 1

    A. Djouadi,The anatomy of electroweak symmetry breaking tome ii: The higgs bosons in the minimal supersymmetric model,Physics Reports459(2008) 1

    work page 2008

  16. [16]

    Fleischer and F

    J. Fleischer and F. Jegerlehner,Radiative corrections to higgs-boson decays in the weinberg-salam model,Phys. Rev. D23(1981) 2001

    work page 1981

  17. [17]

    ATLAS Collaboration,A search for thezγdecay mode of the higgs boson inppcollisions at√s= 13tev with the atlas detector,Phys. Lett. B809(2020) 135754

    work page 2020

  18. [18]

    Tumasyan, others and CMS Collaboration,Search for higgs boson decays to a z boson and a photon in proton-proton collisions at√s= 13tev,JHEP05(2023) 233

    A. Tumasyan, others and CMS Collaboration,Search for higgs boson decays to a z boson and a photon in proton-proton collisions at√s= 13tev,JHEP05(2023) 233

    work page 2023

  19. [19]

    Drell and T.-M

    S.D. Drell and T.-M. Yan,Massive lepton-pair production in hadron-hadron collisions at high energies,Phys. Rev. Lett.25(1966) 316

    work page 1966

  20. [20]

    CMS Collaboration,Measurements of differentialzboson production cross sections in proton–proton collisions at√s= 13tev,JHEP12(2019) 061

    work page 2019

  21. [21]

    ATLAS Collaboration,Measurement of double-differential charged-current drell–yan cross-sections at high transverse masses inppcollisions at√s= 13tev with the atlas detector,JHEP07(2025) 026

    work page 2025

  22. [22]

    Two-body decay kinematics

    H.-C. Cheng, “Two-body decay kinematics.” CERN Indico note, 2015

    work page 2015

  23. [23]

    Sjöstrand, S

    T. Sjöstrand, S. Ask, J.R. Christiansen et al.,An introduction to pythia 8.2,Comput. Phys. Commun.191(2015) 159

    work page 2015

  24. [24]

    Handbook of LHC Higgs Cross Sections: 4. Deciphering the Nature of the Higgs Sector

    D. de Florian et al.,Handbook of lhc higgs cross sections: 4. deciphering the nature of the higgs sector,CERN Yellow Rep. Monogr.2(2017) [1610.07922]

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  25. [25]

    Baldi, P

    P. Baldi, P. Sadowski and D. Whiteson,Searching for exotic particles in high-energy physics with deep learning,Nat. Commun.5(2014) 4308

    work page 2014

  26. [26]

    Machine Learning in High Energy Physics Community White Paper

    K. Albertsson et al.,Machine learning in high energy physics community white paper,arXiv (2018) [1807.02876]

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  27. [27]

    Freund and R.E

    Y. Freund and R.E. Schapire,A decision-theoretic generalization of on-line learning and an application to boosting,Journal of Computer and System Sciences55(1997) 119

    work page 1997

  28. [28]

    Fisher, C

    A. Fisher, C. Rudin and F. Dominici,All models are wrong, but many are useful: Learning a variable’s importance by studying an entire class of prediction models simultaneously,Journal of Machine Learning Research20(2019) 1

    work page 2019

  29. [29]

    Höcker et al.,Tmva — toolkit for multivariate data analysis,PoSACA T(2007) 040

    A. Höcker et al.,Tmva — toolkit for multivariate data analysis,PoSACA T(2007) 040

    work page 2007

  30. [30]

    Looking for a light Higgs boson in the overlooked channel

    J.S. Gainer, W.-Y. Keung, I. Low and P. Schwaller,Looking for a light Higgs boson in the Zγ→ℓℓγchannel,Phys. Rev. D86(2012) 033010 [1112.1405]. – 23 –

    work page internal anchor Pith review Pith/arXiv arXiv 2012

  31. [31]

    Improving the sensitivity of Higgs boson searches in the golden channel

    J.S. Gainer, K. Kumar, I. Low and R. Vega-Morales,Improving the sensitivity of higgs boson searches in the golden channel,JHEP11(2011) 027 [1108.2274]. [37]CMScollaboration,Measurements of properties of the higgs boson decaying into the four-lepton final state in pp collisions at√s= 13tev,JHEP11(2017) 047

    work page internal anchor Pith review Pith/arXiv arXiv 2011

  32. [32]

    Kolmogorov,Sulla determinazione empirica di una legge di distribuzione,Giornale dell’Istituto Italiano degli Attuari4(1933) 83

    A.N. Kolmogorov,Sulla determinazione empirica di una legge di distribuzione,Giornale dell’Istituto Italiano degli Attuari4(1933) 83

    work page 1933

  33. [33]

    Smirnov,Table for estimating the goodness of fit of empirical distributions,Annals of Mathematical Statistics19(1948) 279

    N. Smirnov,Table for estimating the goodness of fit of empirical distributions,Annals of Mathematical Statistics19(1948) 279

    work page 1948

  34. [34]

    XGBoost: A Scalable Tree Boosting System

    T. Chen and C. Guestrin,Xgboost: A scalable tree boosting system,arXiv preprint 1603.02754(2016)

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  35. [35]

    Friedman,Greedy function approximation: A gradient boosting machine,Annals of Statistics29(2001) 1189

    J.H. Friedman,Greedy function approximation: A gradient boosting machine,Annals of Statistics29(2001) 1189

    work page 2001

  36. [36]

    Pedregosa et al.,Scikit-learn: Machine learning in python,Journal of Machine Learning Research12(2011) 2825

    F. Pedregosa et al.,Scikit-learn: Machine learning in python,Journal of Machine Learning Research12(2011) 2825

    work page 2011

  37. [37]

    Cox,The regression analysis of binary sequences,Journal of the Royal Statistical Society: Series B (Methodological)20(1958) 215

    D.R. Cox,The regression analysis of binary sequences,Journal of the Royal Statistical Society: Series B (Methodological)20(1958) 215

    work page 1958

  38. [38]

    Fawcett,An introduction to roc analysis,Pattern Recognition Letters27(2006) 861

    T. Fawcett,An introduction to roc analysis,Pattern Recognition Letters27(2006) 861

    work page 2006

  39. [39]

    Navas et al

    P.D. Group,Review of Particle Physics, vol. 2024 (2024), 10.1103/PhysRevD.110.030001

    work page doi:10.1103/physrevd.110.030001 2024

  40. [40]

    Apollinari, I

    G. Apollinari, I. Béjar Alonso, O. Brüning, M. Lamont and L. Rossi, eds.,High-Luminosity Large Hadron Collider (HL-LHC): Technical Design Report, vol. 4 ofCERN Yellow Reports: Monographs, CERN (2017), 10.23731/CYRM-2017-004. – 24 –

    work page doi:10.23731/cyrm-2017-004 2017