Pith. sign in

REVIEW 4 major objections 5 minor 3 cited by

Hunting and identifying coloured resonances in four top events with machine learning

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Fusing jet images with kinematic variables, one network could give the HL-LHC a 5-sigma discovery reach near 2 TeV for colour-octet and colour-sextet scalars and identify which colour representation produced the excess.

desk verdict Solid pair-production reach and a new sextet projection, but the single-production discrimination claim collapses under the paper's own definitional caveat. read the letter →

arxiv 2506.04318 v1 pith:YMBRTCO5 submitted 2025-06-04 hep-ph hep-ex

Thomas Flacke , Jeong Han Kim , Manuel Kunkel , Jun Seung Pi , Werner Porod This is my paper
classification hep-phhep-ex
keywords colouroctetscalarsextetfour-topfinalstatesame-signdileptonjetimagesconvolutionalneuralnetworkHL-LHCdiscoveryreachflow
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

The paper sets out to show that the high-luminosity LHC could discover, or rule out, pair-produced colour-octet and colour-sextet scalars that decay into top-quark pairs, using a neural network on four-top events with two same-sign leptons. For a centre-of-mass energy of 14 TeV and a luminosity of $3~\text{ab}^{-1}$, it claims a $5\sigma$ discovery reach of $m_8 = 1.8$ TeV and $m_6 = 1.92$ TeV, and exclusion reaches of $2.02$ TeV and $2.14$ TeV, for the octet and sextet, respectively. It further claims that retraining the same architecture lets the network identify which colour representation produced an excess, and whether single production contributed, using only the events a discovery would already supply. The reason this matters is that the colour charge of a new particle cannot be measured directly; the network reads it from colour-flow radiation patterns and event topology that kinematic-only searches leave unused.

What carries the argument

The load-bearing object is the event image: each event is projected onto a $50\times50$ grid in the $\eta$--$\phi$ plane centred on the geometric midpoint of the two same-sign leptons, with artificial periodicity in $\eta$, giving three channels — charged hadrons, neutral hadrons and the isolated leptons — alongside a kinematic set $K$ of invariant masses, angular separations, transverse momenta, missing transverse momentum and $S_T$. A multilayer perceptron consumes $K$ while a convolutional network consumes the images, and the two branches are merged before the output layer, with twenty training runs averaged to estimate systematic uncertainties. The physical mechanism carrying the discrimination is colour flow: different colour representations produce different QCD radiation patterns and lepton angular correlations — sextet events cluster the same-sign leptons near $\phi = \pm\pi/2$, octet events spread them — and the images retain these differences. For the identification step, the networks' output scores are treated as probability densities $f_j(x)$, and a log-likelihood ratio $t = -2\ln(L_1/L_2)$ over the events a discovery would provide separates competing signal hypotheses.

What would settle it

Rerun the same preselection and network training on signal and background samples with roughly 200 pileup collisions overlaid per event: if the neutral-hadron image channel loses its separating power, the quoted discovery masses (1.8 and 1.92 TeV) and the clean separation of signal processes at 1.8 TeV would shrink, directly testing the paper's central assumption. A second, eventual check is observational: with a full data set of $3~\text{ab}^{-1}$, either a four-top excess appears in the same-sign dilepton channel at the projected masses, or a null result excludes pair production up to about 2.0–2.1 TeV.

Watch

Extended reading notes

Core claim

The central claim is that colour-octet and colour-sextet scalars decaying to top-quark pairs, produced either in pairs or singly, can be both found and identified at the HL-LHC through the four-top final state with two same-sign leptons. With a network that fuses a multilayer perceptron on kinematic variables with a convolutional network on $3\times50\times50$ jet images of charged hadrons, neutral hadrons and the two isolated leptons, the paper finds a $5\sigma$ discovery reach of $m_8 = 1.8$ TeV and $m_6 = 1.92$ TeV for pair production at $\sqrt{s}=14$ TeV and $3~\text{ab}^{-1}$, and $95\%$ exclusion reaches of $m_8 = 2.02$ TeV and $m_6 = 2.14$ TeV; including octet single production with a top coupling of $\lambda_8 = 1.1$ extends the reach to higher masses. Retrained to discriminate among signal hypotheses, the same architecture separates octet pair production, octet single production and sextet pair production at a benchmark mass of 1.8 TeV, using a log-likelihood ratio test statistic built from network score distributions, and produces intermediate distributions when two processes are mixed. The paper concludes that the method transfers to other spins and colour structures because it depends only on the four-top final state.

Load-bearing premise

The reach and discrimination numbers assume that the jet images, especially the neutral-particle channel, keep their discriminating power under the HL-LHC's roughly 200 pileup collisions per bunch crossing, because the networks are trained and evaluated on detector simulations without pileup overlays.

Editorial extensions

If this is right

  • If the projected reach holds, the full HL-LHC dataset of $3~\text{ab}^{-1}$ at 14 TeV would discover or exclude pair-produced colour-octet and colour-sextet scalars decaying to tops at masses roughly 0.4–0.6 TeV beyond current limits ($m_8 \leq 1.38$ TeV, $m_6 \leq 1.51$ TeV).
  • An excess at a mass up to about 1.9 TeV could be assigned to octet pair production, sextet pair production, or octet single production using the same events that establish the discovery, because the score distributions for the three hypotheses separate cleanly at $m = 1.8$ TeV.
  • A sizable top coupling ($\lambda_8 = 1.1$) that makes single production dominate over pair production extends the sensitivity to higher masses, so the discovery potential is not capped by the pair-production-only numbers.
  • The neural network and the image data add discriminating power most visibly where kinematics alone are weakest, namely for single production; for pair production the kinematic-only network already matches the imaging network, so the two components are complementary.
  • Because the analysis depends only on the $t\bar{t}t\bar{t}$ final state, the same trained pipeline can be aimed at four-top signals of other spins and colour structures without redesign.

Reading between the lines

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

  • The largest performance gap between the imaging network and the kinematic-only network occurs for single production, which suggests the images carry information about the production mechanism (colour connections to the initial state) rather than just the mass scale; a feature-attribution study of which pixels drive the network would test where that information lives.
  • The paper separates pure signal hypotheses and shows that mixed-process distributions lie in between, but it does not fit the mixture fraction; estimating the amount of single production in an excess is a direct extension that would matter for composite-Higgs scenarios in which octet and sextet states coexist.
  • The pileup-free assumption is the one place the quoted numbers could move; overlaying roughly 200 minimum-bias collisions per event in the detector simulation and repeating the training is a direct check of the 1.8–2.1 TeV frontiers.
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

4 major / 5 minor

Summary. The paper studies the HL-LHC prospects for discovering and identifying color-octet and color-sextet scalar resonances that decay into top pairs, focusing on the same-sign dilepton four-top final state. Signal events are generated with MadGraph5, showered with Pythia8, and passed through a Delphes simulation; a combined MLP+CNN architecture is trained on kinematic variables and jet images to separate signal from Standard Model backgrounds. The authors report expected discovery reaches of m8 = 1.8 TeV and m6 = 1.92 TeV, exclusion reaches of m8 = 2.02 TeV and m6 = 2.14 TeV at 3 ab^-1, and a second-stage network discrimination that can distinguish S8S8, S6S6*, and S8 t tbar production, including a claimed ability to determine whether a significant single-production contribution is present.

Significance. If the reach numbers are taken at face value, the paper provides useful and timely projections for a challenging four-top final state at the HL-LHC. The pair-production octet exclusion is independently corroborated by the recent recast in Ref. [61], which is a strong sanity check, and the ML methodology is generic enough to be applied to other coloured or neutral resonances. The main limitations are the use of leading-order signal cross sections, the absence of a pileup simulation for the jet-image CNN, and a conceptual issue in the definition and statistical handling of the 'single production' component. These issues do not invalidate the pair-production discovery/exclusion analysis, but they do affect one of the headline claims in the abstract.

major comments (4)
  1. [Sec. 5.2 and Appendix A] The abstract and Sec. 5.2 claim that the network 'can determine whether there is a significant contribution from single production to pair production', but Appendix A states that splitting the cross section into single and pair production is 'strictly speaking – unphysical', because the gg-initiated pair-production contribution has the same initial and final state as associated S8 t tbar production and is the same order in all couplings. The 'S8 t tbar' training sample is generated at lambda8 = 1.1 and therefore includes a pair-production component, while the S8S8 sample uses lambda8 << 1. The statement that single production dominates relies on the proxy sigma(4t; S8) - sigma(S8S8), whose validity requires negligible interference, which is not demonstrated. The classifier is therefore effectively separating lambda8 = 1.1 events from lambda8 << 1 events rather than a physically defined single-production rate. Please reframe the claim as a benchmark-coupling discrimination test, or introduce a physical observable (for example an invariant-mass sideband of the second S8) and quantify interference effects, or remove the claim from the abstract.
  2. [Sec. 5.2, Eq. (5.3)] Even if the single/pair definition issue is set aside, the log-likelihood ratio in Eq. (5.3) only tests pure hypotheses f1 versus f2. For the claimed ability to determine whether there is a 'significant contribution from single production', one would need a mixture model with an unknown fraction r and a confidence interval or upper limit on r. The gray 'both' distributions in Fig. 7 are shown qualitatively, but no mixture-based significance or limit is computed. Please provide such an analysis or soften the claim to 'the network can discriminate between the pure benchmark processes'.
  3. [Sec. 4 and Sec. 5.1] The jet-image CNN uses neutral-particle images generated with Delphes but without HL-LHC pileup, although the anticipated average pileup is approximately 200 interactions per bunch crossing. Section 4 acknowledges this and refers to Ref. [29] and previous work [28], but no quantitative estimate is given for the present analysis. Since the CNN advantage over the kinematic-only MLP is largest for the S8 t tbar sample (Fig. 5c and Sec. 5.1), the projected sensitivity for that process and the signal-identification results in Sec. 5.2 could degrade if neutral images are contaminated by pileup. Please either include a pileup-degraded projection (for example by repeating the analysis without the neutral channel or with a pileup overlay) or explicitly state which conclusions are robust independent of the neutral images.
  4. [Sec. 2 and Fig. 6] The signal cross sections are computed at leading order because no NLO-capable UFO is available for the sextet. Leading-order predictions for coloured scalar production can have sizeable scale uncertainties, and the K-factors for octet and sextet pair production may differ, directly affecting the quoted 5-sigma and 95% exclusion reaches. The manuscript should quantify the scale/PDF uncertainty (for example by varying mu_R and mu_F around m_S) and, if feasible, use NLO for the octet while giving an estimate of the sextet K-factor. At minimum, the uncertainty should be displayed in Fig. 6 and reflected in the reach numbers.
minor comments (5)
  1. [Abstract, Sec. 5.1, Sec. 6] The exclusion reach is quoted as m8 = 2.02 TeV and m6 = 2.14 TeV in the abstract and conclusions, but Sec. 5.1 states m8 <= 2.00 TeV and m6 <= 2.12 TeV. These numbers should be harmonized.
  2. [Sec. 2] The sentence 'We use MadGraph5 aMC@NLO v3.5.3 with with NNPDF2.3QED...' contains a duplicated 'with'.
  3. [Fig. 11 and Sec. C.2] The text and figure refer to '600 neutrons' in a network layer; this should be 'neurons'.
  4. [Sec. 4] The kinematic dataset K is used without normalization, while the pixel intensities are also unnormalized. Since MLPs can be sensitive to feature scaling, please state whether standardization was tested or motivate the choice.
  5. [Sec. 5.2] The definition of N5sigma should be made explicit: it is the number of events required for discovery using the signal-versus-background network, not the number of events passing the subsequent signal-discrimination score cut. Please clarify this in the text.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the HL-LHC reach is a self-contained Monte-Carlo projection, externally matched by Ref. [61]; the only caveat is the paper's own admission that the 'single production' label is an operational lambda8=1.1 sample.

full rationale

The paper's central results are derived from its own Monte-Carlo simulations: signal and background events are generated with MadGraph/Pythia/Delphes, the networks are trained and evaluated on separate training/validation/holdout test sets, and the discovery/exclusion reach is obtained by rescaling the simulated signal cross section until Z=5 or Z=1.64. No parameter is fitted to external data, and the projected exclusion reach for the colour octet (about 2.0 TeV) is compared with the independent external result of Ref. [61] (m8 <= 2.04 TeV), which closely agrees. The process-identification study in Sec. 5.2 uses the classifier score distributions as PDFs and constructs a likelihood-ratio test on held-out events with 20 training runs; this is a standard evaluation of classifier separation power, not a double-counted prediction. Self-citations to Refs. [28,29] for the jet-image method and for pileup expectations are not load-bearing in a circular sense, because the present paper retrains and re-evaluates the networks on new simulations and shows ROC curves and limits. The one genuine caveat is Appendix A, which states that splitting the cross section into 'single' and 'pair' production is 'strictly speaking unphysical', and defines the 'single production' sample as the lambda8=1.1 dataset, which also contains pair-production contributions. Consequently, the abstract's claim that the network can 'determine whether there is a significant contribution from single production to pair production' is imprecise: the network is actually distinguishing lambda8=1.1 from lambda8<<1 event samples. This is a well-posedness/interpretation limitation, not a circular reduction of the kind where an output is equivalent to a fitted input by construction, and it does not affect the independent pair-production reach or the colour-representation discrimination. Overall circularity is minimal.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central claim depends on the simplified model parameters (lambda_8 = 1.1, Br = 1), the fidelity of the Monte Carlo simulation, and the pileup assumption for jet images. No new entities are introduced; the S8 and S6 are existing composite-Higgs model states.

free parameters (1)
  • lambda_8 (top coupling of octet scalar) = 1.1
    Chosen by hand so that single production dominates over pair production while keeping the width below 10% of the mass (narrow width approximation). Affects the S8 t tbar signal sample and the corresponding reach.
assumptions (4)
  • domain assumption Br(S8 -> t tbar) = Br(S6 -> tt) = 1
    Decays to top quarks are assumed to dominate, ignoring subleading WZW and lighter-quark decays present in composite Higgs models. This makes the projected reach optimistic for realistic models; stated in Section 2.
  • domain assumption Leading-order cross sections for signal processes are accurate enough for the reach estimate
    The paper uses LO cross sections for both octet and sextet signals, with NLO unavailable for the sextet UFO; a K-factor uncertainty is not propagated to the reach.
  • domain assumption Neutral jet images remain usable at HL-LHC pileup levels
    Pileup (average about 200) is not simulated in this work; the paper relies on prior work [28] showing a weaker but robust result when neutral images are excluded. Section 4.
  • domain assumption The MLP+CNN architecture trained on MC events will generalize to real detector data
    Standard assumption in simulation-based studies; domain shift is not quantified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hunting and identifying coloured resonances in four top events with machine learning." pith.science (2026). https://pith.science/paper/YMBRTCO5

@misc{pith2026250604318,
  author       = {Pith},
  title        = {Pith review of: Hunting and identifying coloured resonances in four top events with machine learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YMBRTCO5}},
  note         = {Machine review of arXiv:2506.04318}
}
abstract

We study prospects to search for pair or singly produced colour octet or colour sextet scalars which decay into two top quarks at the LHC. We focus on the same-sign lepton final state. We train a neural network comprising a simple multilayer perceptron combined with a convolutional neural network to optimize the separation of signal and background events. For LHC operated at 14 TeV and a luminosity of 3 ab$^{-1}$ we find an expected discovery reach of $m_8=1.8$ TeV and $m_6=1.92$ TeV for pair produced colour octets and sextets, respectively, and an expected exclusion reach of $m_8=2.02$ TeV and $m_6=2.14$ TeV. In a second step, we retrain the same network architecture to discriminate between signal processes. The network can clearly distinguish between the different colour representations. Moreover, we can also determine whether there is a significant contribution from single production to pair production for the same final state. The methodology can be applied to BSM candidates of different spin and colour representations.

Discussion (0). Sign in to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Matter Unification and Lepton Flavour Violation

    hep-ph 2026-03 conditional novelty 6.0 of 10

    Muon-to-electron conversion at Mu2e can probe quark–lepton unification scales up to ~10^4 TeV in the minimal SU(4)_C×SU(2)_L×U(1)_R model, even where rare kaon decays are suppressed by unknown mixing angles.

  2. Dark matter in composite Higgs models with a scotogenic EFT

    hep-ph 2026-07 conditional novelty 5.0 of 10

    In the SU(6)/Sp(6) composite Higgs model with a scotogenic Z2, three of four neutral pseudo-Goldstone dark matter candidates can reproduce the observed relic density, while the SU(2)L triplet candidate fails; spin-1 r...

  3. Quark-Lepton Unification Signatures

    hep-ph 2025-11 conditional novelty 5.0 of 10

    In the minimal low-scale quark-lepton unification model, scalar leptoquarks decay mainly to third-generation quarks and leptons, and current LHC measurements exclude masses below roughly 1.0–1.15 TeV in the y2=0 benchmark.

Reference graph

Works this paper leans on

51 extracted references · 16 canonical work pages · cited by 3 Pith papers

  1. [61]

    Boosting Beyond: A Novel Approach to Probing Top-Philic Resonances at the LHC

    L. Darm´ e, B. Fuks, H.-L. Li, M. Maltoni, O. Mattelaer and J. Touch` eque,Novel approach to probing top-philic resonances with boosted four-top tagging,Phys. Rev. D111(2025) 055037, [2404.14482]

    work page Pith review arXiv 2025

  2. [29]

    J. H. Kim, M. Kim, K. Kong, K. T. Matchev and M. Park,Portraying Double Higgs at the Large Hadron Collider,JHEP09(2019) 047, [1904.08549]

    work page Pith review arXiv 2019

  3. [28]

    Flacke, J

    T. Flacke, J. H. Kim, M. Kunkel, P. Ko, J. S. Pi, W. Porod et al.,Uncovering doubly charged scalars with dominant three-body decays using machine learning,JHEP11(2023) 009, [2304.09195]

    arXiv 2023

  4. [1]

    S. P. Martin,A Supersymmetry primer,Adv. Ser. Direct. High Energy Phys.18(1998) 1–98, [hep-ph/9709356]

    arXiv 1998

  5. [2]

    Benakli, M

    K. Benakli, M. Goodsell, F. Staub and W. Porod,Constrained minimal Dirac gaugino supersymmetric standard model,Phys. Rev. D90(2014) 045017, [1403.5122]

    arXiv 2014

  6. [3]

    D. B. Kaplan,Flavor at SSC energies: A New mechanism for dynamically generated fermion masses,Nucl. Phys. B365(1991) 259–278

    1991

  7. [4]

    Ferretti and D

    G. Ferretti and D. Karateev,Fermionic UV completions of Composite Higgs models,JHEP03 (2014) 077, [1312.5330]

    arXiv 2014

  8. [5]

    Ferretti,Gauge theories of Partial Compositeness: Scenarios for Run-II of the LHC,JHEP 06(2016) 107, [1604.06467]

    G. Ferretti,Gauge theories of Partial Compositeness: Scenarios for Run-II of the LHC,JHEP 06(2016) 107, [1604.06467]

    arXiv 2016

Show all 51 references
  1. [6]

    Cacciapaglia, T

    G. Cacciapaglia, T. Flacke, M. Kunkel and W. Porod,Phenomenology of unusual top partners in composite Higgs models,JHEP02(2022) 208, [2112.00019]

    2022 arXiv

  2. [7]

    Cacciapaglia, H

    G. Cacciapaglia, H. Cai, A. Deandrea, T. Flacke, S. J. Lee and A. Parolini,Composite scalars at the LHC: the Higgs, the Sextet and the Octet,JHEP11(2015) 201, [1507.02283]

    2015 arXiv

  3. [8]

    Krause, A

    C. Krause, A. Pich, I. Rosell, J. Santos and J. J. Sanz-Cillero,Colorful Imprints of Heavy States in the Electroweak Effective Theory,JHEP05(2019) 092, [1810.10544]

    2019 arXiv

  4. [9]

    Cacciapaglia, G

    G. Cacciapaglia, G. Ferretti, T. Flacke and H. Serˆ odio,Light scalars in composite Higgs models, Front. in Phys.7(2019) 22, [1902.06890]

    2019 arXiv

  5. [10]

    Cacciapaglia, C

    G. Cacciapaglia, C. Pica and F. Sannino,Fundamental Composite Dynamics: A Review,Phys. Rept.877(2020) 1–70, [2002.04914]

    2020 arXiv

  6. [11]

    Cacciapaglia, A

    G. Cacciapaglia, A. Deandrea, M. Kunkel and W. Porod,Coloured spin-1 states in composite Higgs models,JHEP06(2024) 092, [2404.02198]

    2024 arXiv

  7. [12]

    Dorˇ sner, S

    I. Dorˇ sner, S. Fajfer, A. Greljo, J. F. Kamenik and N. Koˇ snik,Physics of leptoquarks in precision experiments and at particle colliders,Phys. Rept.641(2016) 1–68, [1603.04993]

    2016 arXiv

  8. [13]

    Fileviez Perez and M

    P. Fileviez Perez and M. B. Wise,Low Scale Quark-Lepton Unification,Phys. Rev. D88(2013) 057703, [1307.6213]

    2013 arXiv

  9. [14]

    Faber, M

    T. Faber, M. Hudec, M. Malinsk´ y, P. Meinzinger, W. Porod and F. Staub,A unified leptoquark model confronted with lepton non-universality inB-meson decays,Phys. Lett. B787(2018) 159–166, [1808.05511]

    2018 arXiv

  10. [15]

    Faber, M

    T. Faber, M. Hudec, H. Koleˇ sov´ a, Y. Liu, M. Malinsk´y, W. Porod et al.,Collider phenomenology of a unified leptoquark model,Phys. Rev. D101(2020) 095024, [1812.07592]

    2020 arXiv

  11. [16]

    Fileviez Perez and C

    P. Fileviez Perez and C. Murgui,Flavor anomalies and quark-lepton unification,Phys. Rev. D 106(2022) 035033, [2203.07381]

    2022 arXiv

  12. [17]

    Davoudiasl, J

    H. Davoudiasl, J. L. Hewett and T. G. Rizzo,Bulk gauge fields in the Randall-Sundrum model, Phys. Lett. B473(2000) 43–49, [hep-ph/9911262]. – 18 –

    2000 arXiv

  13. [18]

    Pomarol,Gauge bosons in a five-dimensional theory with localized gravity,Phys

    A. Pomarol,Gauge bosons in a five-dimensional theory with localized gravity,Phys. Lett. B486 (2000) 153–157, [hep-ph/9911294]

    2000 arXiv

  14. [19]

    Chang, J

    S. Chang, J. Hisano, H. Nakano, N. Okada and M. Yamaguchi,Bulk standard model in the Randall-Sundrum background,Phys. Rev. D62(2000) 084025, [hep-ph/9912498]

    2000 arXiv

  15. [20]

    Bajc and G

    B. Bajc and G. Gabadadze,Localization of matter and cosmological constant on a brane in anti-de Sitter space,Phys. Lett. B474(2000) 282–291, [hep-th/9912232]

    2000 arXiv

  16. [21]

    Cheng, K

    H.-C. Cheng, K. T. Matchev and M. Schmaltz,Bosonic supersymmetry? Getting fooled at the CERN LHC,Phys. Rev. D66(2002) 056006, [hep-ph/0205314]

    2002 arXiv

  17. [22]

    Lillie, L

    B. Lillie, L. Randall and L.-T. Wang,The Bulk RS KK-gluon at the LHC,JHEP09(2007) 074, [hep-ph/0701166]

    2007 arXiv

  18. [23]

    J. R. Ellis, V. A. Khoze and W. J. Stirling,Hadronic antenna patterns to distinguish production mechanisms for large E(T) jets,Z. Phys. C75(1997) 287–296, [hep-ph/9608486]

    1997 arXiv

  19. [24]

    Gallicchio and M

    J. Gallicchio and M. D. Schwartz,Seeing in Color: Jet Superstructure,Phys. Rev. Lett.105 (2010) 022001, [1001.5027]

    2010 arXiv

  20. [25]

    A. Atre, R. S. Chivukula, P. Ittisamai and E. H. Simmons,Distinguishing Color-Octet and Color-Singlet Resonances at the Large Hadron Collider,Phys. Rev. D88(2013) 055021, [1306.4715]

    2013 arXiv

  21. [26]

    R. S. Chivukula, P. Ittisamai, K. Mohan and E. H. Simmons,Color discriminant variable and scalar diquarks at the LHC,Phys. Rev. D92(2015) 075020, [1507.06676]

    2015 arXiv

  22. [27]

    T. Han, I. M. Lewis, H. Liu, Z. Liu and X. Wang,A guide to diagnosing colored resonances at hadron colliders,JHEP08(2023) 173, [2306.00079]

    2023 arXiv

  23. [30]

    Huang, S.-b

    L. Huang, S.-b. Kang, J. H. Kim, K. Kong and J. S. Pi,Portraying double Higgs at the Large Hadron Collider II,JHEP08(2022) 114, [2203.11951]

    2022 arXiv

  24. [31]

    Butter et al.,The Machine Learning landscape of top taggers,SciPost Phys.7(2019) 014, [1902.09914]

    A. Butter et al.,The Machine Learning landscape of top taggers,SciPost Phys.7(2019) 014, [1902.09914]

    2019 arXiv

  25. [32]

    Chakraborty, S

    A. Chakraborty, S. H. Lim and M. M. Nojiri,Interpretable deep learning for two-prong jet classification with jet spectra,JHEP07(2019) 135, [1904.02092]. [33]D0collaboration, V. M. Abazov et al.,Measurement of Color Flow int ¯tEvents fromp¯ p Collisions at √s=1.96TeV,Phys. Rev....

    2019 arXiv

  26. [37]

    Belyaev, G

    A. Belyaev, G. Cacciapaglia, H. Cai, G. Ferretti, T. Flacke, A. Parolini et al.,Di-boson signatures as Standard Candles for Partial Compositeness,JHEP01(2017) 094, [1610.06591]

    2017 arXiv

  27. [38]

    Degrande, C

    C. Degrande, C. Duhr, B. Fuks, D. Grellscheid, O. Mattelaer and T. Reiter,UFO - The Universal FeynRules Output,Comput. Phys. Commun.183(2012) 1201–1214, [1108.2040]

    2012 arXiv

  28. [39]

    Alloul, N

    A. Alloul, N. D. Christensen, C. Degrande, C. Duhr and B. Fuks,FeynRules 2.0 - A complete toolbox for tree-level phenomenology,Comput. Phys. Commun.185(2014) 2250–2300, [1310.1921]

    2014 arXiv

  29. [40]

    Fuks and L

    B. Fuks and L. Darme,FeynRules model top-philic resonances, https://feynrules.irmp.ucl.ac.be/wiki/TopHeavyRes

  30. [41]

    Darm´ e, B

    L. Darm´ e, B. Fuks and M. Goodsell,Cornering sgluons with four-top-quark events,Phys. Lett. B784(2018) 223–228, [1805.10835]

    2018 arXiv

  31. [42]

    Darm´ e, B

    L. Darm´ e, B. Fuks and F. Maltoni,Top-philic heavy resonances in four-top final states and their EFT interpretation,JHEP09(2021) 143, [2104.09512]

    2021 arXiv

  32. [43]

    Alwall and C

    J. Alwall and C. Duhr,FeynRules modelSextetDiquarks.fr, https://feynrules.irmp.ucl.ac.be/wiki/Sextets

  33. [44]

    T. Han, I. Lewis and T. McElmurry,QCD Corrections to Scalar Diquark Production at Hadron Colliders,JHEP01(2010) 123, [0909.2666]

    2010 arXiv

  34. [45]

    Alwall, R

    J. Alwall, R. Frederix, S. Frixione, V. Hirschi, F. Maltoni, O. Mattelaer et al.,The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations,JHEP07(2014) 079, [1405.0301]

    2014 arXiv

  35. [46]

    Hirschi and O

    V. Hirschi and O. Mattelaer,Automated event generation for loop-induced processes,JHEP10 (2015) 146, [1507.00020]. [47]NNPDFcollaboration, R. D. Ball, V. Bertone, S. Carrazza, L. Del Debbio, S. Forte, A. Guffanti et al.,Parton distributions with QED corrections,Nucl. Phys. B87...

    2015 arXiv

  36. [48]

    S. Jung, D. Lee and K.-P. Xie,BeyondM t¯t: learning to search for a broadt ¯tresonance at the LHC,Eur. Phys. J. C80(2020) 105, [1906.02810]

    2020 arXiv

  37. [49]

    Deandrea, T

    A. Deandrea, T. Flacke, B. Fuks, L. Panizzi and H.-S. Shao,Single production of vector-like quarks: the effects of large width, interference and NLO corrections,JHEP08(2021) 107, [2105.08745]

    2021 arXiv

  38. [50]

    Sj¨ ostrand, S

    T. Sj¨ ostrand, S. Ask, J. R. Christiansen, R. Corke, N. Desai, P. Ilten et al.,An introduction to PYTHIA 8.2,Comput. Phys. Commun.191(2015) 159–177, [1410.3012]. [51]DELPHES 3collaboration, J. de Favereau, C. Delaere, P. Demin, A. Giammanco, V. Lema ˆ ıtre, A. Mertens et al.,...

    2015 arXiv

  39. [52]

    Cacciari, G

    M. Cacciari, G. P. Salam and G. Soyez,FastJet User Manual,Eur. Phys. J.C72(2012) 1896, [1111.6097]. – 20 – [53]ATLAScollaboration,Technical Design Report for the ATLAS Inner Tracker Pixel Detector, Tech. Rep. ATLAS-TDR-030, 2017. [54]ATLAScollaboration,Expected performance of ...

    2012 arXiv

  40. [55]

    Bertolini, P

    D. Bertolini, P. Harris, M. Low and N. Tran,Pileup Per Particle Identification,JHEP10 (2014) 059, [1407.6013]. [56]CMScollaboration, A. M. Sirunyan et al.,Pileup mitigation at CMS in 13 TeV data,JINST 15(2020) P09018, [2003.00503]

    2014 arXiv

  41. [57]

    Cowan, K

    G. Cowan, K. Cranmer, E. Gross and O. Vitells,Asymptotic formulae for likelihood-based tests of new physics,Eur. Phys. J. C71(2011) 1554, [1007.1727]

    2011 arXiv

  42. [58]

    Kunkel,Collider Phenomenology of Composite Higgs Models, Ph.D

    M. Kunkel,Collider Phenomenology of Composite Higgs Models, Ph.D. thesis, U. Wurzburg (main), 2025. 10.25972/OPUS-39149. [59]ATLAScollaboration, G. Aad et al.,Search for R-parity-violating supersymmetry in a final state containing leptons and many jets with the ATLAS experimen...

    2021 arXiv

  43. [62]

    Glorot, A

    X. Glorot, A. Bordes and Y. Bengio,Deep sparse rectifier neural networks, inProceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics(G. Gordon, D. Dunson and M. Dud ´ ık, eds.), vol. 15 ofProceedings of Machine Learning Research, (Fort L...

    2011

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.