Pith. sign in

REVIEW 3 major objections 5 minor 20 references

Recent results on searches with boosted Higgs bosons at CMS

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Backed by graph-network and transformer jet taggers, CMS reports new dedicated searches in four boosted-Higgs topologies, including first LHC limits on gamma-H production.

desk verdict A clean but unfinished proceedings summary of four real CMS boosted-Higgs analyses; the placeholder in Section 6 is the only hard flaw. read the letter →

arxiv 2507.11977 v1 pith:J5KC6U36 submitted 2025-07-16 hep-ex cs.LG

Farouk Mokhtar This is my paper
classification hep-excs.LG
keywords boostedHiggsbosonsjetsubstructurelarge-radiusjetsmachine-learningtaggersParticleNet-MDGloParTpairproductiongamma-H
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

These proceedings claim that the boosted-Higgs regime at the LHC has become experimentally accessible, and that CMS has now produced dedicated searches in four channels in which a Higgs boson is highly boosted and its decay products merge into a single large-radius jet. The key enablers are two deep-learning taggers: ParticleNet-MD, a graph neural network for $X(qq)$ jets, and GloParT, a transformer covering 37 classes including $X(VV)$ and top pairs. With them, CMS reports an expected significance of $1.76\,\sigma$ for boosted $H\to WW$ (observed 0), a constraint on the quartic coupling modifier $\kappa_{2V}$ from $HH\to bbVV$, $\mu_{VH}=0.72^{+0.75}_{-0.71}$ for $V(qq)H(bb)$, and first LHC constraints on $\gamma H$ production. Several of these are first dedicated searches in their final states; each quantitative result originates in a separate CMS analysis summarized here.

What carries the argument

The machinery that carries the argument is the combination of soft-drop groomed large-radius jets with two learned taggers: ParticleNet-MD, a graph neural network that sorts jets into eight classes including $X(qq)$, and GloParT, a transformer that sorts jets into 37 classes including $X(VV)$ and top pairs. The groomed jet mass $m_{SD}$ is the observable used for signal extraction, and the taggers are trained in simulation, with GloParT calibrated to data using Lund jet plane reweighting (a data-driven correction to simulated jet substructure) and, for the one-lepton $H\to WW$ channel, fine-tuned into a dedicated model $P(H_{1\ell})$ that improved expected significance by nearly 70%. These objects carry the argument by suppressing the QCD multijet background by orders of magnitude at fixed signal efficiency.

What would settle it

Measure the background efficiency of ParticleNet-MD on a data control sample (for example QCD multijet events) at a fixed signal efficiency in the same phase space as Figure 1, such as $450 < p_T < 600$ GeV and $90 < m_{SD} < 140$ GeV; if the data efficiency exceeds the simulation ROC curve by more than the quoted systematic uncertainty, the expected significances such as the $1.76\,\sigma$ for boosted $H\to WW$ would not be reproduced.

Watch

Extended reading notes

Core claim

The central discovery reported is that a dedicated boosted-Higgs search program is now running at CMS, enabled by large-radius jets and modern deep-learning taggers. The paper presents four analyses: $H\to WW$ in all-hadronic and one-lepton categories, $HH\to bbVV$, $V(qq)H(bb)$, and $\gamma H$. Its headline numbers are an observed (expected) significance of 0 ($1.76\,\sigma$) for $H\to WW$; the second-strongest CMS constraint on $\kappa_{2V}$ from $HH\to bbVV$; $\mu_{VH}=0.72^{+0.75}_{-0.71}$ and $\mu_{VZ}=0.09\pm0.63$ for the boosted $VH$ search; and first LHC limits on $\sigma_{\gamma H}$ using $H\to bb$ and $H\to4\ell$ decays. The paper's claim is that these are the first dedicated searches in their final states and that the taggers are what make that sensitivity possible.

Load-bearing premise

The load-bearing premise is that the machine-learning jet taggers, trained on simulated jets, classify real proton-proton jets with the same efficiency and background rejection after the data-driven substructure calibration and control-region extrapolations.

Editorial extensions

If this is right

  • The $H\to WW$ search is statistics-limited: with observed significance 0 and expected $1.76\,\sigma$, adding HL-LHC data is a direct route to first evidence.
  • The $HH\to bbVV$ analysis gives a new, independent bound on the quartic coupling modifier $\kappa_{2V}$ that can be combined with the stronger $HH\to4b$ channel.
  • The first dedicated boosted $V(qq)H(bb)$ search measures $\mu_{VH}=0.72^{+0.75}_{-0.71}$, consistent with the standard model, so no enhancement of vector-boson couplings appears in this channel.
  • The first LHC constraints on $\gamma H$ production, although statistics-limited, open a new production mode where an observation would be clear evidence of beyond-standard-model physics.
  • The planned first ParticleNet $bb$ triggers and improved GloParT taggers are expected to bring significant sensitivity gains in the HL-LHC era.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: The near-70% expected-significance gain from fine-tuning GloParT on the one-lepton $H\to WW$ signature suggests that signal-specific transfer learning may be the fastest near-term sensitivity lever for rare boosted-Higgs final states.
  • Editorial inference: Because GloParT and ParticleNet-MD are generic classifiers of 37 and 8 jet classes, the same trained architectures could be transferred to boosted decays of other heavy resonances (for example, $W$, $Z$, top, or exotic particles) with only modest fine-tuning, making this search program a template for a broader boosted-object program.
  • Editorial inference: The quoted significances and limits assume that tagger performance measured in simulation survives the Lund-plane calibration in data; a public data-vs-simulation closure test of the ROC curves in the exact phase space of Figure 1 would either confirm or undermine that assumption.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. This proceedings contribution from the CMS collaboration summarizes recent searches for boosted Higgs bosons. It describes ML-based large-radius jet tagging with ParticleNet-MD, GloParT, and a fine-tuned GloParT variant; the boosted H->WW search with 0-lepton and 1-lepton categories; the HH->bbVV search and its constraints on the kappa_2V coupling; the V(qq)H(bb) measurement with mu_VH = 0.72 (+0.75, -0.71); and first LHC constraints on gamma-H production and light-quark Yukawa couplings. The numerical results are taken from the cited CMS physics analysis summaries and prior publications; the paper contains no new derivation, fit, or data analysis.

Significance. If the quoted results are correct, they represent genuine experimental progress: the first dedicated CMS boosted H->WW search, a competitive kappa_2V constraint from HH->bbVV, the first dedicated CMS boosted VH->bb search, and the first LHC constraints on gamma-H production. The paper is a compact, well-referenced compilation of these results, and the cited numbers are internally consistent with the corresponding CMS notes. Its principal value is as a proceedings summary; the quantitative support for the claims resides in the cited CMS analyses, not in this manuscript. The paper also has an unfinished placeholder in Section 6, so as submitted it is not a complete article.

major comments (3)
  1. [Section 6, Fig. 3] The two right-hand panels of Fig. 3 contain the literal placeholder text 'Figures we want to show in the paper' in place of the gamma-H cross-section constraints and the light-quark Yukawa coupling constraints. Section 6 refers to these as 'shown in Fig. 3', so the headline claims of the section cannot be checked from the manuscript. Please replace the placeholders with the final plots, or remove the specific claims and refer the reader to CMS-PAS-HIG-23-011.
  2. [Sections 2-4, Fig. 1] The only performance evidence for the taggers that underpin the quoted significances, signal strengths, and limits is a CMS simulation ROC curve. Section 3 states that GloParT is trained in Monte Carlo simulation and calibrated with Lund jet plane reweighting, but the paper provides no data/MC comparison of tagger score distributions, no control-region closure test, and no reference to where these validation studies are documented in the cited CMS notes. Since all quantitative results in Sections 3-5 depend on these taggers, please explicitly point to the validation sections of CMS-PAS-HIG-24-008, CMS-PAS-HIG-24-017, and CMS-PAS-JME-23-001, or add a short validation summary.
  3. [Section 3] The text reports an observed significance of 0 and expected significance of 1.76 sigma for the combined boosted H->WW search, but it does not give the corresponding observed upper limit on the cross section. For a proceedings summary this is a completeness issue, but it is worth stating the observed 95% CL limit from CMS-PAS-HIG-24-008 so that readers can compare the sensitivity with the quoted expected significance.
minor comments (5)
  1. [Section 4] There is a typo in the sentence 'second-strongest CMS constraints on the the quartic VVHH coupling modifier': the word 'the' is repeated.
  2. [Fig. 1 (middle)] The legend of the middle panel lists 'H WW* qq vs. QCD' twice; one of the entries should presumably denote a different final state, such as the 1-lepton or 3-prong topology. Please correct the legend.
  3. [Section 5] The phrase 'the combined mu_VH is measured (expected) to be' is slightly ambiguous because the expected value is a fit result under the SM hypothesis rather than an experimental measurement. Please phrase it as 'observed (expected under the SM) value' for consistency with Section 3.
  4. [Section 6] The claim 'first gamma-H results at the LHC' appears in the text without an immediate citation to CMS-PAS-HIG-23-011; the reference is listed in the bibliography but not cited at the point of the claim.
  5. [General] The paper would benefit from a sentence at the end of the Introduction stating that all results are taken from the cited CMS notes and that the figures are CMS preliminary material, since the body text otherwise reads as if the quoted numbers are derived in this paper.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the proceedings reports CMS results from external analysis notes, with no derived quantity reducing to a fitted input or self-citation chain.

full rationale

The paper is a collaboration proceedings that summarizes results taken from CMS Physics Analysis Summaries and published CMS papers (Refs. [5,7,11,16,17,20]); it contains no derivation chain in which an output is defined in terms of the input it claims to predict. The quoted expected significance of 1.76 sigma for H→WW, the fitted mu_VH = 0.72(+0.75,-0.71), and the kappa_2V and gamma-H constraints are presented as outcomes of the cited CMS analyses, not as predictions fitted from the same numbers shown in these proceedings. The ML taggers (ParticleNet-MD, GloParT, P(H1ℓ)) are trained on Monte Carlo simulation and calibrated with Lund jet plane reweighting, but the paper does not fit a parameter to data and then call the same quantity a prediction; at most it inherits a simulation-to-data calibration assumption from the cited analyses, which is an external validation concern rather than an internal circular step. The literal placeholder 'Figures we want to show in the paper' in Section 6 makes some results unverifiable from this text, but incompleteness is not circularity. Self-citations to CMS notes are the standard way for a collaboration member to report the collaboration's results and are not load-bearing in the sense of replacing independent evidence with a closed loop. Therefore no circular step is exhibited and the score is 0.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

The proceedings introduces no free parameters or invented entities; it reports fitted quantities (mu_VH, limits on kappa_2V, sigma_gammaH) as final results from companion analyses. The load-bearing assumptions are the validity of the simulation-based tagger calibrations and the statistical methodology.

assumptions (2)
  • domain assumption CMS Monte Carlo simulation accurately models jet substructure and tagger response in the boosted Higgs phase space.
    Sections 2 and 3 present tagger ROC curves labeled 'CMS Simulation Preliminary' and state GloParT is trained in MC simulation and calibrated with Lund jet plane reweighting; all quoted significances rely on this.
  • domain assumption Binned maximum likelihood fits and control-region extrapolations correctly separate signal from background.
    Section 3 describes signal extraction via a fit to the soft-drop mass distribution and QCD background estimated by data control-region extrapolation; the paper gives no closure tests or fit validation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Recent results on searches with boosted Higgs bosons at CMS." pith.science (2026). https://pith.science/paper/J5KC6U36

@misc{pith2026250711977,
  author       = {Pith},
  title        = {Pith review of: Recent results on searches with boosted Higgs bosons at CMS},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J5KC6U36}},
  note         = {Machine review of arXiv:2507.11977}
}
read the original abstract

The study of boosted Higgs bosons at the LHC provides a unique window to probe Higgs boson couplings at high energy scales and search for signs of physics beyond the standard model. In these proceedings, we present recent results on boosted Higgs boson searches at the CMS experiment, highlighting innovative reconstruction and tagging techniques that enhance sensitivity in this challenging regime.

Figures

Figures reproduced from arXiv: 2507.11977 by the authors.

Figure 1
Figure 1. Left: Performance of ParticleNet-MD for H → bb tagging [5]. Middle: GloParT for H → WW tagging [11]. Right: Fine-tuned version of GloParT targeting 1ℓ final state topologies [11]. 2 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. (left). The observed significance for the combined result is 0, while the expected significance is 1.76𝜎. This is the first dedicated study of boosted H → WW decays in CMS. 4. Boosted HH → bbVV In the analysis in Ref. [7], large-radius jets are identified using ParticleNet-MD for H → bb and GloParT for H → WW. The mass observable for signal extraction is the H candidate mass after a regression is applied to improve … view at source ↗
Figure 3
Figure 3. Left: Post-fit H boson candidate 𝑚SD distribution in the signal region, summed over all data-taking periods [17]. Middle: Constraints on 𝜎𝛾H using H → bb and H → 4ℓ [20]. Right: Constraints on the Yukawa couplings of light quarks using H → 4ℓ [20]. 7. Summary and outlook The CMS collaboration has extensively studied single Higgs boson (H) and Higgs boson pair (HH) production in boosted topologies across production m… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

20 extracted references · 7 canonical work pages

  1. [1]

    Observation of a new particle in the search for the Standard Model Higgs boson with the ATLAS detector at the LHC

    ATLAS Collaboration, “Observation of a new particle in the search for the Standard Model Higgs boson with the ATLAS detector at the LHC”,Phys. Lett. B 716 (2012) 1, doi:10.1016/j.physletb.2012.08.020, arXiv:1207.7214

    arXiv 2012

  2. [2]

    Observation of a new boson at a mass of 125GeV with the CMS experiment at the LHC

    CMS Collaboration, “Observation of a new boson at a mass of 125GeV with the CMS experiment at the LHC”,Phys. Lett. B 716 (2012) 30, doi:10.1016/j.physletb.2012.08.021, arXiv:1207.7235

    arXiv 2012

  3. [3]

    The CMS experiment at the CERN LHC

    CMS Collaboration, “The CMS experiment at the CERN LHC”,JINST 3 (2008) S08004, doi:10.1088/1748-0221/3/08/S08004

    doi:10.1088/1748-0221/3/08/s08004 2008

  4. [4]

    ParticleNet: Jet tagging via particle clouds

    H. Qu and L. Gouskos, “ParticleNet: Jet tagging via particle clouds”,Phys. Rev. D 101 (2020) 056019,doi:10.1103/PhysRevD.101.056019, arXiv:1902.08570

    arXiv 2020

  5. [5]

    Performanceofheavy-flavourjetidentificationinboostedtopologiesin proton-proton collisions at√𝑠 = 13 TeV

    CMSCollaboration, “Performanceofheavy-flavourjetidentificationinboostedtopologiesin proton-proton collisions at√𝑠 = 13 TeV”, CMS Physics Analysis Summary CMS-PAS-BTV-22-001, 2023

    work page 2023

  6. [6]

    Particle transformer for jet tagging

    H. Qu, C. Li, and S. Qian, “Particle transformer for jet tagging”, inProceedings of the 39th International Conference on Machine Learning, K. Chaudhuri et al., eds., volume 162, p. 18281. 2022.arXiv:2202.03772

    arXiv 2022

  7. [7]

    Search for highly energetic double Higgs boson production in the two bottom quark and two vector boson all-hadronic final state

    CMS Collaboration, “Search for highly energetic double Higgs boson production in the two bottom quark and two vector boson all-hadronic final state”, CMS Physics Analysis Summary CMS-PAS-HIG-23-012, 2024

    2024

  8. [8]

    Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques

    CMS Collaboration, “Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques”,JINST 15(2020) P06005, doi:10.1088/1748-0221/15/06/p06005, arXiv:2004.08262

    arXiv 2020

Show all 20 references
  1. [9]

    Performance of the mass-decorrelated DeepDoubleX classifier for double-b and double-c large-radius jets with the CMS detector

    CMS Collaboration, “Performance of the mass-decorrelated DeepDoubleX classifier for double-b and double-c large-radius jets with the CMS detector”, CMS Detector Performance Note CMS-DP-2022-041, 2022

    2022

  2. [10]

    Identification of heavy-flavour jets with the CMS detector inpp collisions at 13TeV

    CMS Collaboration, “Identification of heavy-flavour jets with the CMS detector inpp collisions at 13TeV”,JINST 13(2018) P05011, doi:10.1088/1748-0221/13/05/P05011, arXiv:1712.07158

    2018 arXiv

  3. [11]

    SearchforHiggsbosonproductionathightransversemomentuminthe WW* decay channel in proton-proton collisions at√𝑠 = 13 TeV

    CMSCollaboration,“SearchforHiggsbosonproductionathightransversemomentuminthe WW* decay channel in proton-proton collisions at√𝑠 = 13 TeV”, CMS Physics Analysis Summary CMS-PAS-HIG-24-008, 2025

    2025

  4. [12]

    A new method for correcting the substructure of multi-prong jets using Lund jet plane reweighting in the CMS experiment

    CMS Collaboration, “A new method for correcting the substructure of multi-prong jets using Lund jet plane reweighting in the CMS experiment”, CMS Physics Analysis Summary CMS-PAS-JME-23-001, 2025. 5 Recent results on searches with boosted Higgs bosons at CMS Farouk Mokhtar

    2025

  5. [13]

    Lund plane reweighting for jet substructure correction

    CMS Collaboration, “Lund plane reweighting for jet substructure correction”, CMS Detector Performance Note CMS-DP-2023-046, 2023

    2023

  6. [14]

    Soft drop

    A. J. Larkoski, S. Marzani, G. Soyez, and J. Thaler, “Soft drop”,JHEP 05 (2014) 146, doi:10.1007/JHEP05(2014)146, arXiv:1402.2657

    2014 arXiv

  7. [15]

    A portrait of the Higgs boson by the CMS experiment ten years after the discovery

    CMS Collaboration, “A portrait of the Higgs boson by the CMS experiment ten years after the discovery”, Nature 607 (2022) 60,doi:10.1038/s41586-022-04892-x, arXiv:2207.00043

    2022 arXiv

  8. [16]

    Search for nonresonant pair production of highly energetic Higgs bosons decaying to bottom quarks

    CMS Collaboration, “Search for nonresonant pair production of highly energetic Higgs bosons decaying to bottom quarks”,Phys. Rev. Lett. 131(2023) 041803, doi:10.1103/PhysRevLett.131.041803, arXiv:2205.06667

    2023 arXiv

  9. [17]

    CMS Collaboration, “Search for a boosted Higgs boson decaying to bottom quark pairs in association with a hadronically decayingWor Zboson with the CMS detector using proton-proton collisions at√𝑠 = 13TeV”, CMS Physics Analysis Summary CMS-PAS-HIG-24-017, 2025

    2025

  10. [18]

    Measurement of boosted Higgs bosons produced via vector boson fusion or gluon fusion in the H→bb decay mode using LHC proton-proton collision data at√𝑠 = 13 TeV

    CMS Collaboration, “Measurement of boosted Higgs bosons produced via vector boson fusion or gluon fusion in the H→bb decay mode using LHC proton-proton collision data at√𝑠 = 13 TeV”,JHEP 12(2024) 035,doi:10.1007/JHEP12(2024)035, arXiv:2407.08012

    2024 arXiv

  11. [19]

    Precise determination of the mass of the Higgs boson and tests of compatibility of its couplings with the standard model predictions using proton collisions at 7 and 8TeV

    CMS Collaboration, “Precise determination of the mass of the Higgs boson and tests of compatibility of its couplings with the standard model predictions using proton collisions at 7 and 8TeV”,Eur. Phys. J. C 75(2015) 212,doi:10.1140/epjc/s10052-015-3351-7, arXiv:1412.8662

    2015 arXiv

  12. [20]

    Search for𝛾H production and constraints on the Yukawa couplings of light quarks to the Higgs boson

    CMS Collaboration, “Search for𝛾H production and constraints on the Yukawa couplings of light quarks to the Higgs boson”, CMS Physics Analysis Summary CMS-PAS-HIG-23-011, 2025. 6

    2025

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.