pith. machine review for the scientific record. sign in

arxiv: 2604.20456 · v1 · submitted 2026-04-22 · ✦ hep-ph · hep-ex

Recognition: unknown

How Invisible: Regressing The Key Model Parameter for Semi-visible Jet Searches

Yin Li , Bingxuan Liu , Jianbin Wang , Jiaqi Xie , Kairong Xu , Ruihan Ye , Zihuan Huang
Authors on Pith no claims yet

Pith reviewed 2026-05-10 00:26 UTC · model grok-4.3

classification ✦ hep-ph hep-ex
keywords semi-visible jetsr_invregressiondark sectorscollider searchesmachine learningphoton associationinvisible fraction
0
0 comments X

The pith

A regression model reconstructs the invisible fraction r_inv in semi-visible jet events using only high-level physics objects and achieves higher precision than analytical methods.

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

The paper develops a machine learning regression technique to determine the value of r_inv, which sets how much of a semi-visible jet's energy goes into invisible dark matter particles. This parameter is central to models of strongly interacting dark sectors at particle colliders. If the method works as described, it would allow experimenters to measure this key quantity directly from data rather than relying on less precise analytical estimates. Such measurements could tighten constraints on dark sector models and improve the reach of searches for these signals. The approach focuses on events where the jet appears alongside an energetic photon, using only reconstructed objects like jets and photons for the input.

Core claim

The central claim is that a regression model trained to predict r_inv from high-level reconstructed objects in photon-associated SVJ events achieves robust performance across varying signal parameters and significantly higher precision than the previously developed analytical method, providing a new way to conduct SVJ searches that can unify s-channel and t-channel productions.

What carries the argument

The regression model that takes high-level physics objects as input to predict the value of r_inv, the invisible fraction of dark hadrons.

Load-bearing premise

The regression trained on Monte Carlo simulations will generalize to real experimental data, and that high-level reconstructed objects alone contain sufficient information to recover r_inv without low-level features or detailed detector response modeling.

What would settle it

Applying the trained model to actual collider data and finding that the reconstructed r_inv values deviate substantially from those expected from the true underlying parameters, or that performance improves markedly when low-level detector information is added.

Figures

Figures reproduced from arXiv: 2604.20456 by Bingxuan Liu, Jianbin Wang, Jiaqi Xie, Kairong Xu, Ruihan Ye, Yin Li, Zihuan Huang.

Figure 1
Figure 1. Figure 1: FIG. 1: Representative Feynman diagrams of the signal processes with an ISR photon [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2: Distributions of the analytically reconstructed [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3: Distributions of [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4: Distributions of [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5: Two-dimensional comparison between [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6: Mean [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7: Comparison between the ML-regressed and analytically reconstructed [PITH_FULL_IMAGE:figures/full_fig_p014_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8: Mean reconstructed [PITH_FULL_IMAGE:figures/full_fig_p014_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9: Regressed [PITH_FULL_IMAGE:figures/full_fig_p016_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: FIG. 10: Left: comparison of the normalized [PITH_FULL_IMAGE:figures/full_fig_p017_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: FIG. 11: Mean response of [PITH_FULL_IMAGE:figures/full_fig_p017_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: FIG. 12: Distributions of [PITH_FULL_IMAGE:figures/full_fig_p019_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: FIG. 13: 80% contour regions in the two-dimensional ( [PITH_FULL_IMAGE:figures/full_fig_p020_13.png] view at source ↗
read the original abstract

Semi-visible jets (SVJs) provide a characteristic collider signature of strongly interacting dark sectors, in which the key model parameter $r_{\mathrm{inv}}$ controls the fraction of dark hadrons decaying to dark matter candidates. In this work, a regression model is developed to reconstruct $r_{\mathrm{inv}}$ in SVJ events produced in association with an energetic photon. The model uses information from high-level physics objects only, and the training procedure is optimized to ensure applicability. The performance is found to be robust against varying signal parameters and $r_{\mathrm{inv}}$ can be reconstructed at a much higher precision, compared to previously developed analytical method. It offers a new approach to conduct SVJ searches that can potentially unify both $s$-channel and $t$-channel productions, enhancing the sensitivities.

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

3 major / 2 minor

Summary. The manuscript develops a regression model using high-level reconstructed physics objects to estimate the invisible fraction parameter r_inv in semi-visible jet events produced in association with an energetic photon. Trained on Monte Carlo simulations, the model is claimed to achieve substantially higher precision than a prior analytical method, to remain robust under variations in signal parameters, and to enable a unified approach to s-channel and t-channel SVJ searches.

Significance. If the regression generalizes reliably beyond simulation, the ability to extract r_inv with improved precision would be a useful addition to dark-sector search strategies at colliders. Because r_inv directly sets the visible-to-invisible energy fraction, a data-driven estimator could tighten signal modeling, improve background discrimination, and increase sensitivity in both production modes.

major comments (3)
  1. Abstract: the claim that r_inv 'can be reconstructed at a much higher precision' is unsupported by any quantitative metrics, resolution figures, bias values, or direct comparison to the analytical method; without these numbers the performance advantage cannot be evaluated.
  2. Training procedure: no description is given of the loss function, network architecture, hyper-parameter optimization, or validation strategy used to ensure applicability; the robustness claim therefore lacks a concrete test (e.g., explicit parameter scans or cross-validation results).
  3. Generalization to data: the central applicability claim rests on an MC-trained mapping that is never tested against simulation-to-data discrepancies (jet energy scale, pile-up, fragmentation, or dark-sector hadronization mismodeling); absent data-driven closure tests or systematic variations, the unification of s- and t-channel searches remains unproven.
minor comments (2)
  1. Abstract: the phrase 'previously developed analytical method' requires a citation.
  2. Notation: ensure r_inv is defined once and used consistently in all equations and figures.

Simulated Author's Rebuttal

3 responses · 1 unresolved

We thank the referee for the constructive comments, which have helped us improve the clarity and completeness of the manuscript. We address each major point below and have made revisions to incorporate additional details and discussions where feasible.

read point-by-point responses

  1. Referee: Abstract: the claim that r_inv 'can be reconstructed at a much higher precision' is unsupported by any quantitative metrics, resolution figures, bias values, or direct comparison to the analytical method; without these numbers the performance advantage cannot be evaluated.

    Authors: We agree that the abstract should provide quantitative support for the performance claim to allow proper evaluation. The body of the manuscript contains the relevant results from our Monte Carlo studies, including resolution, bias, and comparisons. In the revised version, we have updated the abstract to explicitly include key metrics (e.g., the achieved resolution on r_inv and the factor of improvement over the analytical method) drawn from those studies. revision: yes

  2. Referee: Training procedure: no description is given of the loss function, network architecture, hyper-parameter optimization, or validation strategy used to ensure applicability; the robustness claim therefore lacks a concrete test (e.g., explicit parameter scans or cross-validation results).

    Authors: We acknowledge the need for explicit details on the training setup. The revised manuscript now includes a dedicated subsection describing the network architecture, the loss function (mean squared error), the hyper-parameter tuning procedure, and the validation approach, including cross-validation results and explicit scans over signal parameters to substantiate the robustness claim. revision: yes

  3. Referee: Generalization to data: the central applicability claim rests on an MC-trained mapping that is never tested against simulation-to-data discrepancies (jet energy scale, pile-up, fragmentation, or dark-sector hadronization mismodeling); absent data-driven closure tests or systematic variations, the unification of s- and t-channel searches remains unproven.

    Authors: We agree this is a limitation of the current work, which relies entirely on Monte Carlo simulation. In the revised manuscript, we have added a discussion of potential systematic effects from jet energy scale, pile-up, and hadronization variations, along with how they could impact the regression. However, full data-driven closure tests and experimental validation of the unification approach are beyond the scope of this simulation study and would require collaboration with experimental analyses. revision: partial

standing simulated objections not resolved
  • Full data-driven validation and closure tests against real collider data or detailed mismodeling effects cannot be performed within this Monte Carlo-based study.

Circularity Check

0 steps flagged

No circularity: regression performance is empirical evaluation on labeled MC, not a derivation reducing to inputs.

full rationale

The paper trains a supervised regression on Monte Carlo events in which r_inv is an explicit input label used to generate the samples. Reported precision, robustness, and comparison to an analytical method are measured on held-out simulated events; this is standard ML evaluation and does not invoke any self-definitional equations, fitted parameters renamed as predictions, or load-bearing self-citations that would make the claimed reconstruction equivalent to the training inputs by construction. The approach is self-contained against external benchmarks (MC truth) and contains no first-principles derivation chain.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim depends on the fidelity of Monte Carlo simulation to real data and on the sufficiency of high-level objects; no free parameters are explicitly fitted in the abstract, but the regression itself introduces many learned weights.

axioms (1)
  • domain assumption Monte Carlo simulation of SVJ events with photon association accurately represents the detector response and underlying physics for the purpose of training and validation.
    Training and performance claims rest on simulated samples whose agreement with data is assumed rather than demonstrated in the abstract.

pith-pipeline@v0.9.0 · 5451 in / 1212 out tokens · 36690 ms · 2026-05-10T00:26:44.548676+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

35 extracted references · 31 canonical work pages · 4 internal anchors

  1. [1]

    The details of the model can be found in ref

    Algorithm training samples The training samples consider aZ ′ boson produced via ans-channel process, decaying into a pair of dark quarks,χ, i.e.,Z ′ →χ¯χ. The details of the model can be found in ref. [18]. Since this work studies the case with an energetic ISR object, we require at least one ISR photon at the generation level. Processes with an addition...

  2. [2]

    Fors-channel SVJ events, the baseline parameter configuration is chosen to be probVector= 0.75,Λ = 35.1539 GeV, m πD =m ρD = 20 GeV

    Robustness testing samples To validate the robustness of the trained model, additional datasets are generated under different dark sector model assumptions. Fors-channel SVJ events, the baseline parameter configuration is chosen to be probVector= 0.75,Λ = 35.1539 GeV, m πD =m ρD = 20 GeV. Single-parameter variations are then performed for the following qu...

  3. [3]

    As shown in Fig

    Generality testing samples In addition to the resonants-channel production mode, SVJs can also be produced through a non-resonantt-channel production. As shown in Fig. 1, the connection between SM and the dark sector is established by a scalar mediator Φ, which couples a quark in SM to a dark quark through a Yukawa-type interaction. The partonic process i...

    2000

  4. [4]

    Rubin, N

    V.C. Rubin, N. Thonnard and W.K. Ford, Jr.,Rotational properties of 21 SC galaxies with a large range of luminosities and radii, from NGC 4605 /R = 4kpc/ to UGC 2885 /R = 122 kpc/,Astrophys. J.238(1980) 471

    1980

  5. [5]

    The Universal Rotation Curve of Spiral Galaxies: I. the Dark Matter Connection

    M. Persic, P. Salucci and F. Stel,The Universal rotation curve of spiral galaxies: 1. The Dark matter connection,Mon. Not. Roy. Astron. Soc.281(1996) 27 [astro-ph/9506004]

    work page Pith review arXiv 1996

  6. [6]

    A direct empirical proof of the existence of dark matter

    D. Clowe, M. Bradac, A.H. Gonzalez, M. Markevitch, S.W. Randall, C. Jones et al.,A direct empirical proof of the existence of dark matter,Astrophys. J. Lett.648(2006) L109 [astro-ph/0608407]. [4]DEScollaboration,Dark Energy Survey Year 1 Results: Curved-Sky Weak Lensing Mass Map,Mon. Not. Roy. Astron. Soc.475(2018) 3165 [1708.01535]. [5]Planckcollaboratio...

    work page Pith review arXiv 2006

  7. [7]

    Jungman, M

    G. Jungman, M. Kamionkowski and K. Griest,Supersymmetric dark matter,Phys. Rept. 267(1996) 195 [hep-ph/9506380]. [7]ATLAScollaboration,Search for new phenomena in events with an energetic jet and missing transverse momentum inppcollisions at √s=13 TeV with the ATLAS detector, Phys. Rev. D103(2021) 112006 [2102.10874]. [8]CMScollaboration,Search for new pa...

    work page arXiv 1996

  8. [8]

    Strassler and K.M

    M.J. Strassler and K.M. Zurek,Echoes of a hidden valley at hadron colliders,Phys. Lett. B 651(2007) 374 [hep-ph/0604261]

    work page arXiv 2007

  9. [9]

    T. Han, Z. Si, K.M. Zurek and M.J. Strassler,Phenomenology of hidden valleys at hadron colliders,JHEP07(2008) 008 [0712.2041]. 23

    work page arXiv 2008

  10. [10]

    Cohen, M

    T. Cohen, M. Lisanti and H.K. Lou,Semivisible Jets: Dark Matter Undercover at the LHC, Phys. Rev. Lett.115(2015) 171804 [1503.00009]

    work page arXiv 2015

  11. [11]

    Cohen, M

    T. Cohen, M. Lisanti, H.K. Lou and S. Mishra-Sharma,LHC Searches for Dark Sector Showers,JHEP11(2017) 196 [1707.05326]

    work page arXiv 2017

  12. [12]

    Beauchesne, E

    H. Beauchesne, E. Bertuzzo and G. Grilli Di Cortona,Dark matter in Hidden Valley models with stable and unstable light dark mesons,JHEP04(2019) 118 [1809.10152]

    work page arXiv 2019

  13. [13]

    Beauchesne, E

    H. Beauchesne, E. Bertuzzo, G. Grilli Di Cortona and Z. Tabrizi,Collider phenomenology of Hidden Valley mediators of spin 0 or 1/2 with semivisible jets,JHEP08(2018) 030 [1712.07160]

    work page arXiv 2018

  14. [14]

    Albouy et al.,Theory, phenomenology, and experimental avenues for dark showers: a Snowmass 2021 report,Eur

    G. Albouy et al.,Theory, phenomenology, and experimental avenues for dark showers: a Snowmass 2021 report,Eur. Phys. J. C82(2022) 1132 [2203.09503]. [18]CMScollaboration,Search for resonant production of strongly coupled dark matter in proton-proton collisions at 13 TeV,JHEP06(2022) 156 [2112.11125]. [19]ATLAScollaboration,Search for new physics in final ...

    work page arXiv 2021

  15. [15]

    Kar and S

    D. Kar and S. Sinha,Exploring jet substructure in semi-visible jets,SciPost Phys.10(2021) 084 [2007.11597]

    work page arXiv 2021

  16. [16]

    Park and M

    M. Park and M. Zhang,Tagging a jet from a dark sector with Jet-substructures at colliders, Phys. Rev. D100(2019) 115009 [1712.09279]

    work page arXiv 2019

  17. [17]

    Cohen, J

    T. Cohen, J. Doss and M. Freytsis,Jet Substructure from Dark Sector Showers,JHEP09 (2020) 118 [2004.00631]

    work page arXiv 2020

  18. [18]

    Cohen, J

    T. Cohen, J. Roloff and C. Scherb,Dark sector showers in the Lund jet plane,Phys. Rev. D 108(2023) L031501 [2301.07732]

    work page arXiv 2023

  19. [19]

    Faucett, S.-C

    T. Faucett, S.-C. Hsu and D. Whiteson,Learning to identify semi-visible jets,JHEP12 (2022) 132 [2208.10062]

    work page arXiv 2022

  20. [20]

    Favaro, M

    L. Favaro, M. Kr¨ amer, T. Modak, T. Plehn and J. R¨ uschkamp,Semi-visible jets, energy-based models, and self-supervision,SciPost Phys.18(2025) 042 [2312.03067]. 24

    work page arXiv 2025

  21. [21]

    Bernreuther, T

    E. Bernreuther, T. Finke, F. Kahlhoefer, M. Kr¨ amer and A. M¨ uck,Casting a graph net to catch dark showers,SciPost Phys.10(2021) 046 [2006.08639]

    work page arXiv 2021

  22. [22]

    Canelli, A

    F. Canelli, A. de Cosa, L.L. Pottier, J. Niedziela, K. Pedro and M. Pierini,Autoencoders for semivisible jet detection,JHEP02(2022) 074 [2112.02864]

    work page arXiv 2022

  23. [23]

    Bhardwaj, C

    A. Bhardwaj, C. Englert, W. Naskar, V.S. Ngairangbam and M. Spannowsky,Equivariant, safe and sensitive — graph networks for new physics,JHEP07(2024) 245 [2402.12449]

    work page arXiv 2024

  24. [24]

    Pedro and P

    K. Pedro and P. Shyamsundar,Optimal mass variables for semivisible jets,SciPost Phys. Core6(2023) 067 [2303.16253]. [31]ATLAScollaboration,Search for low-mass resonances decaying into two jets and produced in association with a photon or a jet at s=13 TeV with the ATLAS detector,Phys. Rev. D 110(2024) 032002 [2403.08547]

    work page arXiv 2023

  25. [25]

    Liu and K

    B. Liu and K. Pedro,Semi-visible jets + X: illuminating dark showers with radiation,JHEP 12(2024) 105 [2409.04741]

    work page arXiv 2024

  26. [26]

    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, 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]

    work page internal anchor Pith review arXiv 2014

  27. [27]

    An Introduction to PYTHIA 8.2

    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 [1410.3012]. [35]DELPHES 3collaboration,DELPHES 3, A modular framework for fast simulation of a generic collider experiment,JHEP02(2014) 057 [1307.6346]

    work page internal anchor Pith review arXiv 2015

  28. [28]

    The anti-k_t jet clustering algorithm

    M. Cacciari, G.P. Salam and G. Soyez,The anti-k t jet clustering algorithm,JHEP04(2008) 063 [0802.1189]

    work page internal anchor Pith review arXiv 2008

  29. [29]

    FastJet user manual

    M. Cacciari, G.P. Salam and G. Soyez,FastJet User Manual,Eur. Phys. J. C72(2012) 1896 [1111.6097]

    work page internal anchor Pith review arXiv 2012

  30. [30]

    W.S. Cho, K. Choi, Y.G. Kim and C.B. Park,M(T2)-assisted on-shell reconstruction of missing momenta and its application to spin measurement at the LHC,Phys. Rev. D79 (2009) 031701 [0810.4853]

    work page arXiv 2009

  31. [31]

    Choi, J.S

    K. Choi, J.S. Lee and C.B. Park,Measuring the Higgs boson mass with transverse mass variables,Phys. Rev. D82(2010) 113017 [1008.2690]

    work page arXiv 2010

  32. [32]

    Buckley, J

    A. Buckley, J. Butterworth, L. Corpe, C. Doglioni, D. Kar, C. Prat et al.,Probing the sensitivity of semi-visible jets to current LHC measurements using the CONTUR toolkit, 25 2502.11237

    work page arXiv

  33. [33]

    Cazzaniga and A

    C. Cazzaniga and A. de Cosa,Leptons lurking in semi-visible jets at the LHC,Eur. Phys. J. C82(2022) 793 [2206.03909]. [42]CMScollaboration,Search fors-channel production of lepton-enriched semivisible jets in proton-proton collisions at 13 TeV, CMS-PAS-EXO-24-029. [43]CMScollaboration,Search for s-channel production of tau-enriched semivisible jets in pro...

    work page arXiv 2022

  34. [34]

    D. Kar, W. Nzuza and S. Sinha,2B or not 2B, a study of bottom-quark-philic semi-visible jets,SciPost Phys. Core7(2024) 071 [2207.01885]

    work page arXiv 2024

  35. [35]

    Carrasco, S

    J. Carrasco, S. Kulkarni, W. Liu, J. Lockyer and J. Zurita,Semi-visible emerging jets, 2511.02918. 26

    work page arXiv