pith. sign in

arxiv: 2606.00695 · v1 · pith:NZ25P7OWnew · submitted 2026-05-30 · ✦ hep-ph · hep-ex

Machine Learning Enhanced Detection of Higgs Chain Decays in Vector Boson Fusion

Shreecheta Chowdhury , Amit Chakraborty , Stefano Moretti This is my paper

Pith reviewed 2026-06-28 18:25 UTC · model grok-4.3

classification ✦ hep-ph hep-ex
keywords Higgs bosonVector boson fusionNMSSMDeep learningCalorimeter dataFour b-jetsBeyond Standard ModelStatistical significance
0
0 comments X

The pith

Deep learning applied to low-level calorimeter data can detect a heavy Higgs decaying to four bottom quarks in vector boson fusion with 4.5 sigma significance at 300 fb inverse.

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

The paper studies a Next-to-Minimal Supersymmetric Standard Model scenario in which a heavy CP-even Higgs boson is produced via vector boson fusion and decays through a chain of two lighter Higgs bosons into four bottom quarks, accompanied by two forward or backward light quarks. It shows that advanced deep learning techniques operating directly on calorimeter information can separate this signal from backgrounds to reach approximately 4.5 sigma statistical significance at an integrated luminosity of 300 fb inverse at the LHC. The chosen beyond-Standard-Model process serves as an example to illustrate how machine learning on raw detector data can support searches for new physics in multi-jet final states.

Core claim

In the NMSSM a heavy CP-even Higgs h2 is produced in the vector boson fusion process qq to qq h2 and decays via the chain h2 to h1 h1 to b b-bar b b-bar; a deep learning classifier trained on simulated events and using only low-level calorimeter information separates the signal from background sufficiently well to yield a statistical significance of approximately 4.5 sigma for an integrated luminosity of 300 fb inverse at the LHC.

What carries the argument

Deep learning model applied to low-level calorimeter information to classify events containing two forward quarks and four b-jets from the h2 chain decay.

If this is right

  • The described deep learning approach can achieve a statistical significance of approximately 4.5 sigma for the h2 to h1 h1 to four b-jets channel at 300 fb inverse.
  • Only low-level calorimeter information is required; no high-level reconstructed objects are needed for the classification.
  • The same methodology provides an illustrative template for searching other beyond-Standard-Model final states that produce multiple b-jets plus forward quarks.
  • The analysis targets the specific NMSSM parameter space where the heavy Higgs is produced in vector boson fusion and decays through the stated chain.

Where Pith is reading between the lines

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

  • The calorimeter-only deep learning strategy could be tested on other Higgs pair production modes or different beyond-Standard-Model decay topologies that yield similar jet multiplicities.
  • Performance differences between simulation and data would need to be quantified with control regions before the method is used in a real search.
  • At higher integrated luminosities expected from future LHC runs the same classifier architecture might yield correspondingly higher significance for the same signal process.

Load-bearing premise

A deep learning model trained on simulated events will retain its reported performance when applied to real collider data without large discrepancies from imperfect detector modeling or background processes.

What would settle it

Apply the trained classifier to actual LHC collision data corresponding to 300 fb inverse and measure whether the observed excess reaches or exceeds 4.5 sigma significance.

Figures

Figures reproduced from arXiv: 2606.00695 by Amit Chakraborty, Shreecheta Chowdhury, Stefano Moretti.

Figure 1
Figure 1. Figure 1: Normalised distributions of the pseudorapidity separation [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Schematic illustration of the DL architecture employed in this work for classifying signal [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The explicit representation of the image branch mentioned in Figure [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The average jet image of signal events for the leading (top left) and sub-leading (top [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: The average jet-image of signal events for the leading (top left) and subleading (top [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: The kinematic branch of Figure [PITH_FULL_IMAGE:figures/full_fig_p015_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Normalised distributions for the invariant mass of the two [PITH_FULL_IMAGE:figures/full_fig_p016_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Normalised distributions for the energy fraction of the leading (left panel) and sub-leading [PITH_FULL_IMAGE:figures/full_fig_p016_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Normalised distributions for the angular separation ( [PITH_FULL_IMAGE:figures/full_fig_p016_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: The ROC curves for both fixed- and variable- [PITH_FULL_IMAGE:figures/full_fig_p018_10.png] view at source ↗
read the original abstract

Over the years, Vector Boson Fusion (VBF) has established itself as one of the most robust production channels for studying the Higgs boson, while also serving as a promising pathway for exploring potential signatures of physics Beyond the Standard Model (BSM) at the Large Hadron Collider (LHC). Following the discovery of a SM-like Higgs boson, new opportunities have arisen to also investigate heavy resonances that decay into SM-like Higgs boson pairs, $hh$, thereby offering valuable insights into the structure of the Higgs sector and the dynamics governing Electro-Weak Symmetry Breaking (EWSB). In this work, we analyze a final state involving, alongside 2 forward/backward light quarks, 4 $b$-quarks emerging from the chain decay $h_2\to h_1h_1\to b\bar b b\bar b$ wherein the heavy CP-even Higgs state $h_2$ is produced in the VBF process $qq\to qqh_2$ and belongs to the Next-to-Minimal Supersymmetric SM (NMSSM). This BSM scenario is used as an illustrative example of the potential of using only low-level calorimeter information enhanced by advanced Deep Learning (DL) methodologies in searching for this channel, which can achieve a statistical significance of approximately $4.5\sigma$, for an integrated luminosity of 300 fb$^{-1}$ at the CERN machine.

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

2 major / 0 minor

Summary. The manuscript proposes a deep learning classifier trained on low-level calorimeter cell information to enhance sensitivity to vector boson fusion production of a heavy CP-even Higgs h2 in the NMSSM, with the chain decay h2 → h1 h1 → bbbb. It reports a statistical significance of approximately 4.5σ for an integrated luminosity of 300 fb^{-1} at 13 TeV, using this BSM scenario as an illustrative example of raw-detector DL methods.

Significance. If the performance on Monte Carlo samples translates to real data, the approach could provide a useful demonstration of how low-level calorimeter inputs combined with modern DL can improve reach for resonant Higgs-pair production in VBF channels, particularly in BSM scenarios with enhanced couplings.

major comments (2)
  1. [Abstract] Abstract: the quoted 4.5σ significance is obtained exclusively from Monte Carlo samples of signal and SM backgrounds; no information is supplied on background modeling, systematic uncertainties, cross-validation strategy, or data selection criteria, making it impossible to assess whether the quoted significance is supported by the analysis.
  2. [Abstract] The central claim relies on a calorimeter-cell DL classifier whose output is evaluated only on simulated events; no control-region comparison, data-driven background estimation, or systematic variation of detector response (jet energy scale, shower modeling, underlying event) is reported, leaving the sim-to-real gap unquantified and directly load-bearing for any real-data significance claim.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our Monte Carlo study. The work uses simulated events to demonstrate the potential of a deep learning classifier on low-level calorimeter information for a BSM VBF Higgs chain decay search, with the NMSSM scenario serving as an illustrative example. We will revise the abstract to more explicitly state the Monte Carlo scope and add clarifying details on methodology.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the quoted 4.5σ significance is obtained exclusively from Monte Carlo samples of signal and SM backgrounds; no information is supplied on background modeling, systematic uncertainties, cross-validation strategy, or data selection criteria, making it impossible to assess whether the quoted significance is supported by the analysis.

    Authors: The quoted significance is computed from Monte Carlo samples of signal and SM backgrounds generated with standard tools. Background modeling follows conventional LHC simulation practices for the relevant processes. The DL classifier employs cross-validation during training, with event selection based on VBF topology and b-jet requirements as described in the methods. We will revise the abstract to state that the significance is obtained from Monte Carlo simulations and ensure the main text supplies the requested details on modeling, selection, and validation strategy. revision: yes

  2. Referee: [Abstract] The central claim relies on a calorimeter-cell DL classifier whose output is evaluated only on simulated events; no control-region comparison, data-driven background estimation, or systematic variation of detector response (jet energy scale, shower modeling, underlying event) is reported, leaving the sim-to-real gap unquantified and directly load-bearing for any real-data significance claim.

    Authors: This is a simulation-based demonstration study and does not claim results on real data; hence no control regions, data-driven background estimation, or detector systematic variations are included. The focus is the improvement achievable with low-level calorimeter inputs in Monte Carlo. We will revise the abstract to emphasize the Monte Carlo nature of the results and include a statement noting that the sim-to-real gap remains unquantified in this work. revision: partial

Circularity Check

0 steps flagged

No circularity: significance is direct MC evaluation

full rationale

The paper applies a DL classifier to Monte Carlo samples of a specific NMSSM VBF signal and SM backgrounds, reporting a 4.5σ significance at 300 fb^{-1} as the output of that evaluation. This is a standard sensitivity projection on simulated data with no mathematical derivation chain, no fitted parameters renamed as predictions, and no self-citation load-bearing steps. The abstract and described method contain no equations or reductions that equate outputs to inputs by construction; the result is the classifier performance on the input samples, which is self-contained and externally falsifiable via future data.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract contains no explicit free parameters, axioms, or invented entities; assessment is limited by lack of full text.

pith-pipeline@v0.9.1-grok · 5781 in / 1134 out tokens · 23988 ms · 2026-06-28T18:25:12.828254+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

46 extracted references · 2 canonical work pages · 2 internal anchors

  1. [1]

    Observation of a new particle in the search for the standard model higgs boson with the atlas detector at the lhc.Physics Letters B, 716(1):1–29, 2012

    G Aad, B Abbott, J Abdallah, et al. Observation of a new particle in the search for the standard model higgs boson with the atlas detector at the lhc.Physics Letters B, 716(1):1–29, 2012

    2012

  2. [2]

    Observation of a new boson at a mass of 125 gev with the cms experiment at the lhc.Physics Letters B, 716(1):30–61, 2012

    S Chatrchyan, V Khachatryan, AM Sirunyan, et al. Observation of a new boson at a mass of 125 gev with the cms experiment at the lhc.Physics Letters B, 716(1):30–61, 2012

    2012

  3. [3]

    Supersymmetry, supergravity and particle physics.Physics Reports, 110(1– 2):1–162, 1984

    Hans Peter Nilles. Supersymmetry, supergravity and particle physics.Physics Reports, 110(1– 2):1–162, 1984

    1984

  4. [4]

    Joseph D. Lykken. Introduction to supersymmetry.arXiv preprint hep-th/9612114, 1996

    work page internal anchor Pith review Pith/arXiv arXiv 1996

  5. [5]

    Princeton University Press, Princeton, NJ, second edition, 1992

    Julius Wess and Jonathan Bagger.Supersymmetry and Supergravity. Princeton University Press, Princeton, NJ, second edition, 1992

    1992

  6. [6]

    Daniel J. H. Chung, Lisa L. Everett, Gordon L. Kane, Steve F. King, Joseph D. Lykken, and Lian-Tao Wang. The soft supersymmetry-breaking lagrangian: Theory and applications. Physics Reports, 407(1–3):1–203, 2005

    2005

  7. [7]

    Haber and Gordon L

    Howard E. Haber and Gordon L. Kane. The search for supersymmetry: Probing physics beyond the standard model.Physics Reports, 117(2–4):75–263, 1985

    1985

  8. [8]

    Stephen P. Martin. A supersymmetry primer. In Gordon L. Kane, editor,Perspectives on Supersymmetry, volume 18 ofAdvanced Series on Directions in High Energy Physics, pages 1–98. World Scientific, 1998. Updated version available as arXiv:hep-ph/9709356

    work page internal anchor Pith review Pith/arXiv arXiv 1998

  9. [9]

    Godbole, and Probir Roy.Theory and Phenomenology of Sparti- cles: An Account of Four-Dimensional N=1 Supersymmetry in High Energy Physics

    Manuel Drees, Rohini M. Godbole, and Probir Roy.Theory and Phenomenology of Sparti- cles: An Account of Four-Dimensional N=1 Supersymmetry in High Energy Physics. World Scientific, Singapore, 2005. 19

    2005

  10. [10]

    Cambridge University Press, Cambridge, England, 2006

    Howard Baer and Xerxes Tata.Weak Scale Supersymmetry: From Superfields to Scattering Events. Cambridge University Press, Cambridge, England, 2006

    2006

  11. [11]

    The next-to-minimal supersymmetric extension of the standard model re- viewed.International Journal of Modern Physics A, 25(18):3505–3602, 2010

    Markos Maniatis. The next-to-minimal supersymmetric extension of the standard model re- viewed.International Journal of Modern Physics A, 25(18):3505–3602, 2010

    2010

  12. [12]

    Teixeira

    Ulrich Ellwanger, Cyril Hugonie, and Ana M. Teixeira. The next-to-minimal supersymmetric standard model.Physics Reports, 496(1–2):1–77, 2010

    2010

  13. [13]

    CRC Press, 2019

    Stefano Moretti and Shaaban Khalil.Supersymmetry Beyond Minimality: From Theory to Experiment. CRC Press, 2019

    2019

  14. [14]

    Dominant production of heavier Higgs bosons through vector boson fusion in the NMSSM.Phys

    Debottam Das. Dominant production of heavier Higgs bosons through vector boson fusion in the NMSSM.Phys. Rev. D, 99(9):095035, 2019

    2019

  15. [15]

    Gianotti et al

    F. Gianotti et al. Physics potential and experimental challenges of the LHC luminosity upgrade. Eur. Phys. J. C, 39:293–333, 2005

    2005

  16. [16]

    Fat b-jet analyses using old and new clustering algorithms in new Higgs boson searches at the LHC.Eur

    Amit Chakraborty, Srinandan Dasmahapatra, Henry Day-Hall, Billy Ford, Shubhani Jain, and Stefano Moretti. Fat b-jet analyses using old and new clustering algorithms in new Higgs boson searches at the LHC.Eur. Phys. J. C, 83(4):347, 2023

    2023

  17. [17]

    Hammad, S

    A. Hammad, S. Moretti, and M. Nojiri. Multi-scale cross-attention transformer encoder for event classification.JHEP, 03:144, 2024

    2024

  18. [18]

    Hammad, Raymundo Ramos, Amit Chakraborty, Pyungwon Ko, and Stefano Moretti

    A. Hammad, Raymundo Ramos, Amit Chakraborty, Pyungwon Ko, and Stefano Moretti. Ex- plainingdataexcessesovertheNMSSMparameterspacewithDeepLearningtechniques.JHEP, 02:077, 2026

    2026

  19. [19]

    Supergauge Invariant Extension of the Higgs Mechanism and a Model for the electron and Its Neutrino.Nucl

    Pierre Fayet. Supergauge Invariant Extension of the Higgs Mechanism and a Model for the electron and Its Neutrino.Nucl. Phys. B, 90:104–124, 1975

    1975

  20. [20]

    A Simple Solution to the Strong CP Problem with a Harmless Axion.Phys

    Michael Dine, Willy Fischler, and Mark Srednicki. A Simple Solution to the Strong CP Problem with a Harmless Axion.Phys. Lett. B, 104:199–202, 1981

    1981

  21. [21]

    Srednicki, and D

    Hans Peter Nilles, M. Srednicki, and D. Wyler. Weak Interaction Breakdown Induced by Supergravity.Phys. Lett. B, 120:346, 1983

    1983

  22. [22]

    J. M. Frere, D. R. T. Jones, and S. Raby. Fermion Masses and Induction of the Weak Scale by Supergravity.Nucl. Phys. B, 222:11–19, 1983

    1983

  23. [23]

    J. P. Derendinger and Carlos A. Savoy. Quantum Effects and SU(2) x U(1) Breaking in Supergravity Gauge Theories.Nucl. Phys. B, 237:307–328, 1984

    1984

  24. [24]

    Dedes, C

    A. Dedes, C. Hugonie, S. Moretti, and K. Tamvakis. Phenomenology of a new minimal super- symmetric extension of the standard model.Phys. Rev. D, 63:055009, 2001

    2001

  25. [25]

    Panagiotakopoulos and A

    C. Panagiotakopoulos and A. Pilaftsis. Higgs scalars in the minimal nonminimal supersym- metric standard model.Phys. Rev. D, 63:055003, 2001

    2001

  26. [26]

    R. D. Peccei and Helen R. Quinn. CP Conservation in the Presence of Instantons.Phys. Rev. Lett., 38:1440–1443, 1977. 20

    1977

  27. [27]

    R. D. Peccei and Helen R. Quinn. Constraints Imposed by CP Conservation in the Presence of Instantons.Phys. Rev. D, 16:1791–1797, 1977

    1977

  28. [28]

    Ya. B. Zeldovich, I. Yu. Kobzarev, and L. B. Okun. Cosmological Consequences of the Spon- taneous Breakdown of Discrete Symmetry.Zh. Eksp. Teor. Fiz., 67:3–11, 1974

    1974

  29. [29]

    Panagiotakopoulos and K

    C. Panagiotakopoulos and K. Tamvakis. Stabilized NMSSM without domain walls.Phys. Lett. B, 446:224–227, 1999

    1999

  30. [30]

    Gunion, and Cyril Hugonie

    Ulrich Ellwanger, John F. Gunion, and Cyril Hugonie. NMHDECAY: A Fortran code for the Higgs masses, couplings and decay widths in the NMSSM.JHEP, 02:066, 2005

    2005

  31. [31]

    NMSPEC: A Fortran code for the sparticle and Higgs masses in the NMSSM with GUT scale boundary conditions.Comput

    Ulrich Ellwanger and Cyril Hugonie. NMSPEC: A Fortran code for the sparticle and Higgs masses in the NMSSM with GUT scale boundary conditions.Comput. Phys. Commun., 177:399–407, 2007

    2007

  32. [32]

    NMHDECAY 2.0: An Updated program for sparticle masses, Higgs masses, couplings and decay widths in the NMSSM.Comput

    Ulrich Ellwanger and Cyril Hugonie. NMHDECAY 2.0: An Updated program for sparticle masses, Higgs masses, couplings and decay widths in the NMSSM.Comput. Phys. Commun., 175:290–303, 2006

    2006

  33. [33]

    Teixeira

    Debottam Das, Ulrich Ellwanger, and Ana M. Teixeira. NMSDECAY: A Fortran Code for Su- persymmetric Particle Decays in the Next-to-Minimal Supersymmetric Standard Model.Com- put. Phys. Commun., 183:774–779, 2012

    2012

  34. [34]

    The automated computation of tree-level and next- to-leading order differential cross sections, and their matching to parton shower simulations

    J Alwall, R Frederix, S Frixione, et al. The automated computation of tree-level and next- to-leading order differential cross sections, and their matching to parton shower simulations. Journal of High Energy Physics, 2014(7):79, 2014

    2014

  35. [35]

    An introduction to pythia 8.2

    Torbjorn Sjöstrand, Stefan Ask, Jesper R Christiansen, et al. An introduction to pythia 8.2. Computer Physics Communications, 191:159–177, 2015

    2015

  36. [36]

    Delphes 3: a modular framework for fast simulation of a generic collider experiment.Journal of High Energy Physics, 2014(2):57, 2014

    J de Favereau, C Delaere, P Demin, et al. Delphes 3: a modular framework for fast simulation of a generic collider experiment.Journal of High Energy Physics, 2014(2):57, 2014

    2014

  37. [37]

    Soft-collinear radiation in nlo processes.Journal of High Energy Physics, 2015(5):123, 2015

    Matteo Cacciari, Jonathan R Gaunt, Gavin P Salam, and Giulia Zanderighi. Soft-collinear radiation in nlo processes.Journal of High Energy Physics, 2015(5):123, 2015

    2015

  38. [38]

    Fastjet user manual.Journal of High Energy Physics, 2010(4):49, 2010

    Matteo Cacciari, Gavin P Salam, and Sebastian Sapeta. Fastjet user manual.Journal of High Energy Physics, 2010(4):49, 2010

    2010

  39. [39]

    Optimisation and performance studies of the ATLASb-tagging algorithms for the 2017-18 LHC run. 7 2017

    2017

  40. [40]

    Search for theHH→b ¯bb¯bprocess via vector-boson fusion production using proton-proton collisions at√s= 13TeV with the ATLAS detector.JHEP, 07:108, 2020

    Georges Aad et al. Search for theHH→b ¯bb¯bprocess via vector-boson fusion production using proton-proton collisions at√s= 13TeV with the ATLAS detector.JHEP, 07:108, 2020. [Erratum: JHEP 01, 145 (2021), Erratum: JHEP 05, 207 (2021)]

    2020

  41. [41]

    Higgs pair production in vector-boson fusion at the LHC and beyond.Eur

    Fady Bishara, Roberto Contino, and Juan Rojo. Higgs pair production in vector-boson fusion at the LHC and beyond.Eur. Phys. J. C, 77(7):481, 2017

    2017

  42. [42]

    Asymptotic formulae for likelihood- based tests of new physics.Eur

    Glen Cowan, Kyle Cranmer, Eilam Gross, and Ofer Vitells. Asymptotic formulae for likelihood- based tests of new physics.Eur. Phys. J. C, 71:1554, 2011. [Erratum: Eur.Phys.J.C 73, 2501 (2013)]. 21

    2011

  43. [43]

    Going deeper with convolutions

    Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, and Andrew Rabinovich. Going deeper with convolutions. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 1–9. IEEE, 2015

    2015

  44. [44]

    Rethinking the inception architecture for computer vision

    Christian Szegedy, Vincent Vanhoucke, Sergey Ioffe, Jonathon Shlens, and Zbigniew Wojna. Rethinking the inception architecture for computer vision. InProceedings of the IEEE confer- ence on computer vision and pattern recognition, pages 2818–2826. IEEE, 2016

    2016

  45. [45]

    Dolan, and Nina Rajcic

    James Barnard, Edmund Noel Dawe, Matthew J. Dolan, and Nina Rajcic. Parton Shower Uncertainties in Jet Substructure Analyses with Deep Neural Networks.Phys. Rev. D, 95(1):014018, 2017

    2017

  46. [46]

    Jets with Variable R.JHEP, 06:059, 2009

    David Krohn, Jesse Thaler, and Lian-Tao Wang. Jets with Variable R.JHEP, 06:059, 2009. 22 6 Appendix 6.1 Architecture and Network Design The analysis employs a multi-layer DL NN architecture specifically designed to exploit complemen- tary information from jet images and high-level kinematic observables (Figure 2). The combined dataset of signal and backg...

    2009