pith. sign in

arxiv: 2506.11783 · v4 · submitted 2025-06-13 · ✦ hep-ex

Learning from all particles in high-energy collisions

Yongfeng Zhu , Yuexin Wang , Hao Liang , Yuzhi Che , Hengyu Wang , Chen Zhou , Huilin Qu , Manqi Ruan This is my paper

Pith reviewed 2026-05-19 10:10 UTC · model grok-4.3

classification ✦ hep-ex
keywords Higgs bosondeep learninghigh-energy collisionsparticle parentagesignal-background separationcollider data analysisAI methods in particle physics
0
0 comments X

The pith

Deep learning on every reconstructed particle in collisions boosts Higgs measurement precision by up to six times.

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

The paper introduces a holistic approach and Advanced Color Singlet Identification that use all particles in a collider event instead of selected subsets. Deep learning infers the parent particles for each reconstructed track or cluster to improve signal-background separation. If these methods work as described, they would deliver substantially sharper results on standard Higgs properties and make previously inaccessible rare Higgs decays detectable with current data. A general reader would care because the work shows how existing collider datasets can yield more physics information through better use of particle-level details and modern machine learning.

Core claim

By leveraging all reconstructed particles and inferring their parentage via deep learning, the holistic approach and Advanced Color Singlet Identification improve the precision of key Higgs physics benchmark measurements by up to sixfold and enable realistic prospects for observing rare Higgs decays previously deemed inaccessible.

What carries the argument

The holistic approach combined with Advanced Color Singlet Identification, which applies deep learning to assign parent origins to every particle in an event.

If this is right

  • Standard Higgs property measurements reach up to six times better precision.
  • Rare Higgs decay channels move from inaccessible to potentially observable.
  • Signal-background discrimination improves in events with many reconstructed particles.
  • Existing high-energy collision datasets can support new physics measurements without additional data taking.

Where Pith is reading between the lines

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

  • The parentage-inference technique could be applied to other rare processes such as top-quark or electroweak measurements in the same datasets.
  • If the simulation-to-data transfer holds, analysis strategies at the High-Luminosity LHC might shift toward full-event deep learning rather than hand-crafted selections.
  • Similar methods might help isolate signals in searches for new particles that also produce complex multi-particle final states.

Load-bearing premise

Models trained on simulated events can infer particle parentage in real collision data without introducing biases large enough to erase the claimed gains in precision.

What would settle it

A side-by-side comparison on actual collider data showing whether the new methods achieve the stated factor-of-six improvement in Higgs measurement uncertainty compared with conventional techniques.

Figures

Figures reproduced from arXiv: 2506.11783 by Chen Zhou, Hao Liang, Hengyu Wang, Huilin Qu, Manqi Ruan, Yongfeng Zhu, Yuexin Wang, Yuzhi Che.

Figure 1
Figure 1. Figure 1: FIG. 1: For an [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2: Signal likelihood distributions for [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3: A comprehensive summary of signal strength [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4: The [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5: The event selection procedure with cut-based [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7: For the three scenarios—holistic, holistic with [PITH_FULL_IMAGE:figures/full_fig_p008_7.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6: With equal statistics of fully-hadronic [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
read the original abstract

Particle colliders stand as an irreplaceable pillar of inquiry for exploring the fundamental building blocks of matter and forces of the Universe, yet fully decoding complex collision event information remains a significant challenge. Recent advances in artificial intelligence (AI) have revolutionized complex data analysis across scientific disciplines, inspiring novel strategies to extract the rich information embedded in collider events. Here we introduce two complementary concepts -- the holistic approach and Advanced Color Singlet Identification -- to enhance signal-background separation, which is a critical prerequisite for precise physics measurements. By leveraging all reconstructed particles and inferring their parentage via deep learning, these methods improve the precision of key Higgs physics benchmark measurements by up to sixfold and enable realistic prospects for observing rare Higgs decays previously deemed inaccessible. Our results demonstrate how integrating particle-level information with modern AI technologies can substantially boost the discovery potential of high-energy colliders, paving a new path to unravel the fundamental physical laws underlying particle physics experiments.

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 / 2 minor

Summary. The manuscript introduces two complementary methods—the holistic approach and Advanced Color Singlet Identification—that leverage all reconstructed particles in collider events and use deep learning to infer their parentage. These techniques are presented as improving signal-background separation, yielding up to sixfold gains in precision for key Higgs physics benchmarks and opening prospects for observing rare Higgs decays previously considered inaccessible.

Significance. If the claimed precision improvements are robustly validated, the work could meaningfully enhance the physics reach of existing and future colliders by extracting more information from the full particle content of events. The integration of particle-level data with modern AI methods addresses a long-standing challenge in complex event reconstruction and could influence analysis strategies for rare processes.

major comments (2)
  1. Abstract: The central claim of up to sixfold precision gains on Higgs benchmarks is stated without any quantitative validation, error budgets, training details, baseline comparisons, or propagation of parentage misassignment rates into final uncertainties. This absence prevents assessment of whether the reported improvements are supported by the data or models.
  2. The manuscript relies on deep-learning parentage inference trained exclusively on simulated events; no explicit validation or quantitative bound is provided on sim-to-real transfer biases arising from detector modeling, pile-up, or reconstruction inefficiencies, which directly affect the claimed signal-background separation and precision gains.
minor comments (2)
  1. Clarify the precise definitions and algorithmic differences between the 'holistic approach' and 'Advanced Color Singlet Identification' early in the text, including any shared or distinct network architectures.
  2. Provide explicit references or citations for the specific Higgs benchmark measurements used to quantify the sixfold improvement.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their careful reading and constructive comments. We address each major comment below and describe the revisions made to strengthen the manuscript.

read point-by-point responses

  1. Referee: Abstract: The central claim of up to sixfold precision gains on Higgs benchmarks is stated without any quantitative validation, error budgets, training details, baseline comparisons, or propagation of parentage misassignment rates into final uncertainties. This absence prevents assessment of whether the reported improvements are supported by the data or models.

    Authors: We agree that the abstract would benefit from additional context. The full quantitative validations, error budgets, training procedures, baseline comparisons, and propagation of misassignment uncertainties are presented in detail in Sections 3–5 of the manuscript. We have revised the abstract to include a concise reference to these supporting analyses and the validation framework. revision: yes

  2. Referee: The manuscript relies on deep-learning parentage inference trained exclusively on simulated events; no explicit validation or quantitative bound is provided on sim-to-real transfer biases arising from detector modeling, pile-up, or reconstruction inefficiencies, which directly affect the claimed signal-background separation and precision gains.

    Authors: We acknowledge the importance of quantifying sim-to-real transfer effects. In the revised manuscript we have added a dedicated subsection that reports systematic studies varying detector modeling, pile-up conditions, and reconstruction efficiencies. These studies provide quantitative bounds on the resulting impact to signal-background separation. While ground-truth parentage labels exist only in simulation, we include closure tests and data-driven cross-checks to constrain the biases. revision: partial

Circularity Check

0 steps flagged

No circularity in empirical DL application to particle parentage inference

full rationale

The paper introduces holistic and Advanced Color Singlet Identification approaches that apply deep learning to infer parentage from all reconstructed particles in collider events. The reported precision gains on Higgs benchmarks are presented as empirical outcomes of this ML-based separation on data, without any visible equations, derivations, or first-principles results that reduce to fitted parameters or self-definitions by construction. No load-bearing self-citations, ansatzes smuggled via prior work, or renaming of known results appear in the abstract or described claims. The central results rest on the performance of models trained on simulation when applied to the target measurements, making the derivation chain self-contained against external benchmarks rather than tautological.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review yields no identifiable free parameters, axioms, or invented entities; the central claim rests on the unstated performance of unspecified deep learning models.

pith-pipeline@v0.9.0 · 5702 in / 1069 out tokens · 51706 ms · 2026-05-19T10:10:47.785600+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

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

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

Reference graph

Works this paper leans on

38 extracted references · 38 canonical work pages · 10 internal anchors

  1. [1]

    The European Strategy Group, Deliberation document on the 2020 Update of the European Strategy for Particle Physics, Tech. Rep. (Geneva, 2020)

    work page 2020

  2. [2]

    Cheng et al., The Physics potential of the CEPC

    H. Cheng et al., The Physics potential of the CEPC. Pre- pared for the US Snowmass Community Planning Exer- cise (Snowmass 2021) (2022), arXiv:2205.08553 [hep-ph]

    work page arXiv 2021

  3. [3]

    The International Linear Collider Technical Design Report - Volume 1: Executive Summary

    T. Behnke, J. E. Brau, B. Foster, J. Fuster, M. Harrison, J. M. Paterson, M. Peskin, M. Stanitzki, N. Walker, and H. Yamamoto, (2013), arXiv:1306.6327 [physics.acc-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2013

  4. [4]

    Abada et al

    A. Abada et al. (FCC), Eur. Phys. J. ST228, 261 (2019)

    work page 2019

  5. [5]

    Brunner et al., (2022), arXiv:2203.09186 [physics.acc- ph]

    O. Brunner et al., (2022), arXiv:2203.09186 [physics.acc- ph]

    work page arXiv 2022

  6. [6]

    Bai et al., in Snowmass 2021 (2021) arXiv:2110.15800 [hep-ex]

    M. Bai et al., in Snowmass 2021 (2021) arXiv:2110.15800 [hep-ex]

    work page arXiv 2021

  7. [7]

    Br¨ uning, H

    O. Br¨ uning, H. Burkhardt, and S. Myers, Progress in Particle and Nuclear Physics 67, 705 (2012)

    work page 2012

  8. [8]

    Evans, Annual Review of Nuclear and Particle Science 61, 435 (2011)

    L. Evans, Annual Review of Nuclear and Particle Science 61, 435 (2011)

    work page 2011

  9. [9]

    Marciano and H

    W. Marciano and H. Pagels, Physics Reports 36, 137 (1978)

    work page 1978

  10. [10]

    Precision Higgs Physics at CEPC

    F. An et al. , Chin. Phys. C 43, 043002 (2019), arXiv:1810.09037 [hep-ex]

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  11. [11]

    Davidson, S

    S. Davidson, S. Hannestad, and G. Raffelt, Journal of High Energy Physics 2000, 003 (2000)

    work page 2000

  12. [12]

    Magill, R

    G. Magill, R. Plestid, M. Pospelov, and Y.-D. Tsai, Phys. Rev. Lett. 122, 071801 (2019)

    work page 2019

  13. [13]

    Liu and Y

    Z. Liu and Y. Zhang, Phys. Rev. D 99, 015004 (2019)

    work page 2019

  14. [14]

    Arbor, a new approach of the Particle Flow Algorithm

    M. Ruan and H. Videau, in International Conference on Calorimetry for the High Energy Frontier (2013) pp. 316– 324, arXiv:1403.4784 [physics.ins-det]

    work page internal anchor Pith review Pith/arXiv arXiv 2013

  15. [15]

    Duarte-Campderros, G

    J. Duarte-Campderros, G. Perez, M. Schlaffer, and A. Soffer, Phys. Rev. D 101, 115005 (2020), arXiv:1811.09636 [hep-ph]

    work page arXiv 2020

  16. [16]

    J. F. Kamenik, A. Korajac, M. Szewc, M. Tammaro, and J. Zupan, Flavor violating Higgs and Z decays at FCC-ee (2023), arXiv:2306.17520 [hep-ph]

    work page arXiv 2023

  17. [17]

    Albert et al., Strange quark as a probe for new physics in the higgs sector (2022), arXiv:2203.07535 [hep-ex]

    A. Albert et al., Strange quark as a probe for new physics in the higgs sector (2022), arXiv:2203.07535 [hep-ex]

    work page arXiv 2022

  18. [18]

    Quark flavour-violating Higgs decays at the ILC

    D. Barducci and A. J. Helmboldt, JHEP 12, 105, arXiv:1710.06657 [hep-ph]

    work page internal anchor Pith review Pith/arXiv arXiv

  19. [19]

    Liang, Y

    H. Liang, Y. Zhu, Y. Wang, Y. Che, M. Ruan, C. Zhou, and H. Qu, Phys. Rev. Lett. 132, 221802 (2024)

    work page 2024

  20. [20]

    Zhu and M

    Y. Zhu and M. Ruan, The European Physical Journal C 79, 1 (2019)

    work page 2019

  21. [21]

    Y. Zhu, H. Cui, and M. Ruan, JHEP 11, 100, arXiv:2203.01469 [hep-ex]

    work page arXiv

  22. [22]

    Y. Wang, H. Liang, Y. Zhu, Y. Che, X. Xia, H. Qu, C. Zhou, X. Zhuang, and M. Ruan, Computer Physics Communications , 109661 (2025)

    work page 2025

  23. [23]

    The CEPC Study Group, CEPC conceptual design report: Volume 2 - physics & detector (2018), arXiv:1811.10545 [hep-ex]

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  24. [24]

    WHIZARD: Simulating Multi-Particle Processes at LHC and ILC

    W. Kilian, T. Ohl, and J. Reuter, Eur. Phys. J. C 71, 1742 (2011), arXiv:0708.4233 [hep-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2011

  25. [25]

    PYTHIA 6.4 Physics and Manual

    T. Sjostrand, S. Mrenna, and P. Z. Skands, JHEP 05, 026, arXiv:hep-ph/0603175

    work page internal anchor Pith review Pith/arXiv arXiv

  26. [26]

    Herwig++ Physics and Manual

    M. Bahr et al. , Eur. Phys. J. C 58, 639 (2008), arXiv:0803.0883 [hep-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2008

  27. [27]

    Herwig 7.0 / Herwig++ 3.0 Release Note

    J. Bellm et al. , Eur. Phys. J. C 76, 196 (2016), arXiv:1512.01178 [hep-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  28. [28]

    A comprehensive guide to the physics and usage of PYTHIA 8.3

    C. Bierlich et al. , SciPost Phys. Codeb. 2022, 8 (2022), arXiv:2203.11601 [hep-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  29. [29]

    Qu and L

    H. Qu and L. Gouskos, Phys. Rev. D 101, 056019 (2020), arXiv:1902.08570 [hep-ph]

    work page arXiv 2020

  30. [30]

    Z. Wu, S. Pan, F. Chen, G. Long, C. Zhang, and P. S. Yu, IEEE Transactions on Neural Networks and Learning Systems 32, 4 (2021)

    work page 2021

  31. [31]

    H. Qu, C. Li, and S. Qian, in Proceedings of the 39th In- ternational Conference on Machine Learning (2022) pp. 18281–18292, arXiv:2202.03772 [hep-ph]

    work page arXiv 2022

  32. [32]

    Vaswani, N

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. u. Kaiser, and I. Polosukhin, in Advances in Neural Information Processing Systems , Vol. 30, edited by I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett (Curran Associates, Inc., 2017)

    work page 2017

  33. [33]

    Drucker and C

    H. Drucker and C. Cortes, in Advances in Neural Infor- mation Processing Systems , Vol. 8, edited by D. Touret- zky, M. Mozer, and M. Hasselmo (MIT Press, 1995)

    work page 1995

  34. [34]

    Z.-X. Chen, Y. Yang, M.-Q. Ruan, D.-Y. Wang, G. Li, S. Jin, and Y. Ban, Chinese Physics C 41, 023003 (2017)

    work page 2017

  35. [35]

    Bai, C.-H

    Y. Bai, C.-H. Chen, Y.-Q. Fang, G. Li, M.-Q. Ruan, J.- Y. Shi, B. Wang, P.-Y. Kong, B.-Y. Lan, and Z.-F. Liu, Chinese Physics C 44, 013001 (2020)

    work page 2020

  36. [36]

    B. R. Webber, in Summer School on Hadronic As- pects of Collider Physics (1994) pp. 49–77, arXiv:hep- ph/9411384

    work page arXiv 1994

  37. [37]

    Duarte-Campderros, G

    J. Duarte-Campderros, G. Perez, M. Schlaffer, and A. Soffer, Physical Review D 101, 115005 (2020)

    work page 2020

  38. [38]

    Altmann et al., ECF A Higgs, electroweak, and top Fac- tory Study , CERN Yellow Reports: Monographs, Vol

    J. Altmann et al., ECF A Higgs, electroweak, and top Fac- tory Study , CERN Yellow Reports: Monographs, Vol. 5/2025 (2025) arXiv:2506.15390 [hep-ex]

    work page arXiv 2025