Pith. sign in

REVIEW 2 major objections 6 minor 48 references

A mixed-representation network reconstructs antineutron direction and momentum from ECAL deposits alone, improving angular precision by up to 96% and reaching ~17% momentum resolution at 1 GeV/c.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.5

2026-07-14 06:46 UTC pith:JPN6A57H

load-bearing objection Real-data antineutron kinematics from ECAL alone, with a clean mixed visual-sequential design and solid held-out tests; the main soft spot is recoil-label precision, not circularity. the 2 major comments →

arxiv 2607.11139 v1 pith:JPN6A57H submitted 2026-07-13 hep-ex

Antineutron reconstruction in electromagnetic calorimeters with mixed-representation learning

Yangu Li , Hongtian Yu , Yuyang Huang , Zhi Cao , Yunxuan Song , Yunfan Liu , Yajun Mao , YangHeng Zheng
show 2 more authors
Xiao-Rui Lyu Qixiang Ye
This is my paper
classification hep-ex
keywords antineutron reconstructionelectromagnetic calorimetermixed-representation learningobject detectionmomentum resolutionBESIIIneutral hadronshadronic showers
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Long-lived neutral hadrons such as antineutrons leave sparse, fragmented energy deposits in electromagnetic calorimeters that were never designed for hadronic showers, so conventional clustering cannot recover their direction or momentum. This paper shows that those deposits still encode usable kinematic information once they are represented in two complementary ways: a visual branch that encodes the averaged centrality-and-diffusion pattern of many events, and a sequential branch that models non-local correlations among the sparse hits of a single event. Their fusion, MrCAL, jointly predicts particle identity, incident direction, and momentum magnitude from real BESIII collision data. Directional precision improves by up to 96% relative to the standard clustering baseline, and momentum magnitude becomes measurable for the first time from ECAL alone, with roughly 17% resolution at 1 GeV/c. The gains hold across held-out decay channels and multi-particle backgrounds, so existing calorimeter systems can support exclusive final-state analyses that previously required partial reconstruction.

Core claim

Antineutron energy deposits in an ECAL exhibit two reproducible regularities—one sparse and non-local in individual events, one central and diffusion-like when averaged—and a mixed visual-plus-sequential network that exploits both can jointly recover particle identity, momentum direction (up to 96% better angular precision than conventional clustering), and momentum magnitude (~17% resolution at 1 GeV/c) from ECAL readouts alone, with stable performance on real data across many physics processes.

What carries the argument

MrCAL (Mixed-representation Calorimetric Network): a unified object-detection architecture that fuses a diffusion-operator visual branch (encoding averaged centrality-and-diffusion priors) with a transformer sequential branch (modeling long-range hit-token correlations) and reconciles their independent candidates into joint predictions of identity, direction, and momentum.

Load-bearing premise

The recoil four-momenta obtained by constraining the other reconstructed particles must be accurate and unbiased enough to serve as ground-truth labels for both direction and magnitude; residual label error would cap the reported resolutions and inflate gains over clustering.

What would settle it

Re-evaluate the same trained models on an independent, high-purity antineutron sample whose four-momenta are measured by a method that does not rely on recoil against the remaining final-state particles (for example, a dedicated tagged beam or a detector with a functioning hadronic calorimeter) and check whether the claimed angular improvement and 17% momentum resolution survive.

Watch this falsifier — get emailed when new claim-graph text bears on it.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 6 minor

Summary. The manuscript introduces MrCAL, a mixed-representation network that combines a visual branch (image rasterization of ECAL cells with a physics-inspired diffusion operator and RetinaNet-style detection) and a sequential branch (transformer over hit tokens with position-cloze pretraining) to reconstruct antineutrons from BESIII ECAL readouts alone. Trained and evaluated on a large real-data corpus (5.2M selected events across 36 exclusive channels, with a momentum-balanced 1.1M-event training/test split and 13 held-out channels), the method jointly predicts particle identity, incident direction, and momentum magnitude. Relative to conventional clustering, directional precision (1/mAB) improves by up to 96% at full efficiency in the single-particle setting; momentum magnitude is regressed for the first time from ECAL alone at ~17% resolution near 1 GeV/c. Multi-particle hybrid and pure-data tests, photon reconstruction, event-level background rejection, and invariant-mass peaks of mother particles are used to argue robustness and physics utility.

Significance. If the reported resolutions and generalization hold under deployment, the work removes a long-standing algorithmic bottleneck for long-lived neutral hadrons in GeV-scale ECALs without hardware changes. The use of a large real-data benchmark (rather than simulation-only training), held-out channels, efficiency–precision curves, and a conventional clustering baseline are genuine strengths. Enabling direct momentum-magnitude measurement from ECAL alone, and recovering clear hyperon/charmonium mass peaks with predicted antineutron four-momenta, would expand the physics reach of BESIII and similar legacy calorimeters. Code availability is noted; the mixed-representation design is a concrete, transferable contribution beyond a pure architecture paper.

major comments (2)
  1. [Sec. 3, Sec. 4] Sec. 3 and Sec. 4 (multi-particle setting): Multi-particle training and the nominal multi-particle benchmarks use a hybrid construction (real antineutron ECAL pattern + two Geant4 photon patterns), while the multi-particle generalization study uses pure real channels that already contain an antineutron plus two photons. The paper does not quantify the domain gap between hybrid and pure multi-particle inputs (e.g., side-by-side mAB/mRE or mis-ID rates on the same pure channels when the model is trained only on hybrid data). Without that comparison, the multi-particle resolution and PID numbers in Fig. 2(c–e) remain harder to interpret as fully data-driven.
  2. [Sec. 3, Fig. 2(a)] Sec. 3 (labels) and Fig. 2(a): Recoil-derived direction and magnitude labels are the sole ground truth. The manuscript correctly notes that both learning methods and the conventional baseline approach a common low-efficiency angular floor set by cell granularity plus residual label uncertainty, but it does not estimate the absolute contribution of recoil kinematic uncertainty (or selection bias from exclusive channel cuts) to the quoted full-efficiency mAB or to the ~17% momentum resolution at 1 GeV/c. A short quantitative bound—e.g., from MC closure or from varying the recoil constraint—would clarify how much of the absolute resolution is detector/shower-limited versus label-limited, and would strengthen the claim that the 96% relative gain is not inflated by label noise.
minor comments (6)
  1. [Title page] ArXiv ePrint is still listed as xxxx.xxxxx on the title page; replace with 2607.11139 for the final version.
  2. [Fig. 1, Introduction] Fig. 1 caption and body: the phrase “penetrating high energy antineutrons” is slightly misleading for the sub-GeV to ~1.5 GeV/c range studied; “high-energy” relative to nuclear thresholds would be clearer.
  3. [Appendix A.1] Appendix A.1: the diffusion coefficient k and time t are described as parameterized, but no values, ranges, or ablation of the diffusion operator versus a plain Swin/FPN backbone are given. A short ablation would help readers judge how load-bearing the heat-conduction prior is.
  4. [Table 1, Fig. 3] Table 1 and Sec. 5: channel labels in Fig. 3 use compact particle notation that is hard to match to Table 1 at a glance; a consistent channel ID column would help.
  5. [Fig. 6] Fig. 6 caption incorrectly says “of the antineutron” in the photon multi-particle panel description; it should refer to the photon.
  6. [Data availability] Data-availability statement is appropriately cautious for a collaboration dataset; consider stating more explicitly which derived quantities (if any) can be shared under BESIII policy, given that code is already public.

Circularity Check

0 steps flagged

No circularity: supervised reconstruction from ECAL hits is evaluated against independent recoil labels and a conventional baseline; claimed gains and momentum regression are empirical, not forced by construction.

full rationale

The paper trains MrCAL (visual + sequential branches under an object-detection interface) on real BESIII ECAL hit patterns to jointly predict identity, direction and magnitude. Supervision labels are obtained by kinematic recoil against independently reconstructed charged/photon final states in exclusive channels (Sec. 3); the network never receives the recoil four-momentum as an input feature. Performance metrics (mAB, mAE/mRE, efficiency–precision curves, mis-ID rates) are measured against those external labels and against a conventional clustering baseline that uses the same ECAL data. The reported 96 % directional-precision gain and ~17 % momentum resolution at 1 GeV/c are therefore empirical outcomes of the learned mapping, not algebraic identities or re-fitted parameters renamed as predictions. Held-out channels (13 of 28) and multi-particle hybrid tests further demonstrate that the model does not merely memorize channel-specific label artifacts. The diffusion operator (App. A.1) cites prior work with author overlap, but the citation supplies only a technical building block and is not invoked as a uniqueness theorem that forces the central claims. No equation reduces a claimed resolution or improvement to an input by construction; the derivation chain is self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

4 free parameters · 3 axioms · 1 invented entities

The central empirical claims rest on standard detector physics, the recoil-labeling procedure, and a set of architectural and training choices. No new physical entities are postulated; free parameters are the usual network hyper-parameters and fusion thresholds.

free parameters (4)
  • diffusion coefficient k and diffusion time t in visual branch
    Parameterized rather than fixed; learned or tuned to encode the averaged centrality-and-diffusion prior (Appendix A.1).
  • pseudo-box size (10× effective cell size)
    Chosen from a 2-D Gaussian fit to averaged images to supply contextual scale for RetinaNet-style supervision.
  • sequence length L=640, energy/angle bin counts, masking ratio 20%
    Architectural choices that map variable hit multiplicities onto fixed transformer inputs and define the position-cloze pre-training task.
  • branch temperature factors T_b,c and angular matching/fusion thresholds
    Calibrate and reconcile candidate detections from the two branches at inference time (Appendix A.3).
axioms (3)
  • domain assumption Recoil four-momenta against reconstructed companion particles provide unbiased ground-truth labels for antineutron direction and magnitude.
    Stated in Sec. 3; all supervised losses and reported resolutions are measured against these labels.
  • domain assumption Geant4 photon showers are sufficiently faithful that hybrid real-antineutron + MC-photon events can be used for multi-particle training.
    Sec. 3; pure real multi-particle channels are reserved only for generalization tests.
  • ad hoc to paper Averaged antineutron energy deposits obey a radially symmetric diffusion-like profile that can be encoded by a heat-conduction operator.
    Empirical regularity observed in Fig. 1 and formalized in Eqs. (A.1)–(A.4).
invented entities (1)
  • Mixed-representation Calorimetric Network (MrCAL) no independent evidence
    purpose: Joint visual-sequential architecture that fuses diffusion-based image features with hit-token self-attention to reconstruct antineutron identity, direction and momentum.
    The paper’s central technical contribution; no independent experimental existence outside this work.

pith-pipeline@v1.1.0-grok45 · 27546 in / 2581 out tokens · 27808 ms · 2026-07-14T06:46:59.614498+00:00 · methodology

0 comments
read the original abstract

A long-standing bottleneck in GeV-scale accelerator experiments lies in reconstructing long-lived neutral hadrons in conventional electromagnetic calorimeters (ECALs), where hadron--nucleus interactions fall outside the detector's native response regime. In this paper, we develop a physics-inspired representation approach for antineutron reconstruction using a large corpus of real collision data. Motivated by two distinct energy deposition patterns from the penetrating high energy antineutrons in ECALs, we propose a Mixed-representation Calorimetric Network (MrCAL) that integrates complementary visual and sequential representation branches within a unified object-detection architecture. This architecture jointly predicts particle identity, momentum direction, and momentum magnitude. Our approach improves the precision of antineutron momentum-direction reconstruction by up to 96% and, for the first time, enables direct measurement of momentum magnitude solely from ECAL readouts, achieving a momentum resolution of approximately 17% at 1 GeV/c. The model maintains robust performance through comprehensive generalization tests spanning a wide variety of physics processes and background environments. This work unlocks a novel measurement capability for legacy ECAL systems at large experimental facilities, broadening their scientific scope via innovative final-state neutral-hadron detection.

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

48 extracted references · 30 linked inside Pith

  1. [1]

    Fr¨ uhwirth and A

    R. Fr¨ uhwirth and A. Strandlie,Pattern Recognition, Tracking and Vertex Reconstruction in Particle Detectors, Particle Acceleration and Detection, Springer (11, 2020), 10.1007/978-3-030-65771-0

    doi:10.1007/978-3-030-65771-0 2020

  2. [2]

    Yussup, M.I

    N. Yussup, M.I. Idris and I. Yusuff,A decade of trends and progress in methods and hardware accelerators for particle track reconstruction: a systematic review,Radiation Physics and Chemistry239(2026) 113377. [3]BaBarcollaboration,The BaBar detector,Nucl. Instrum. Meth. A479(2002) 1 [hep-ex/0105044]. [4]CLEOcollaboration,The CLEO-II detector,Nucl. Instrum...

    Pith/arXiv arXiv 2026

  3. [3]

    Achasov et al.,STCF conceptual design report (Volume 1): Physics & detector,Front

    M. Achasov et al.,STCF conceptual design report (Volume 1): Physics & detector,Front. Phys. (Beijing)19(2024) 14701 [2303.15790]

    arXiv 2024

  4. [4]

    Pietropaolo et al.,Neutron detection techniques fromµeV to GeV,Phys

    A. Pietropaolo et al.,Neutron detection techniques fromµeV to GeV,Phys. Rept.875(2020) 1

    2020

  5. [5]

    He,Simulation and reconstruction of the BESIII EMC,J

    M. He,Simulation and reconstruction of the BESIII EMC,J. Phys. Conf. Ser.293(2011) 012025. [10]BESIIIcollaboration,Measurement of the Absolute Branching Fraction and Decay Asymmetry ofΛ→nγ,Phys. Rev. Lett.129(2022) 212002 [2206.10791]. [11]BESIIIcollaboration,Study of decay dynamics andCPasymmetry inD + →K 0 Le+νe decay, Phys. Rev. D92(2015) 112008 [1510....

    arXiv 2011

  6. [6]

    Scoville et al.,The Cosmic Evolution Survey (COSMOS): Overview,Astrophys

    N. Scoville et al.,The Cosmic Evolution Survey (COSMOS): Overview,Astrophys. J. Suppl. 172(2007) 1 [astro-ph/0612305]

    Pith/arXiv arXiv 2007

  7. [7]

    Da Cunha, S

    E. Da Cunha, S. Charlot and D. Elbaz,A simple model to interpret the ultraviolet, optical and infrared emission from galaxies,Monthly Notices of the Royal Astronomical Society388 (2008) 1595 [https://academic.oup.com/mnras/article-pdf/388/4/1595/3036508/mnras0388-1595.pdf]

    2008

  8. [8]

    Paganini, L

    M. Paganini, L. de Oliveira and B. Nachman,CaloGAN : Simulating 3D high energy particle showers in multilayer electromagnetic calorimeters with generative adversarial networks, Phys. Rev. D97(2018) 014021 [1712.10321]

    Pith/arXiv arXiv 2018

  9. [9]

    Chekalina, E

    V. Chekalina, E. Orlova, F. Ratnikov, D. Ulyanov, A. Ustyuzhanin and E. Zakharov, Generative Models for Fast Calorimeter Simulation: the LHCb case,EPJ Web Conf.214 (2019) 02034 [1812.01319]. – 23 –

    Pith/arXiv arXiv 2019

  10. [10]

    Belayneh et al.,Calorimetry with deep learning: particle simulation and reconstruction for collider physics,Eur

    D. Belayneh et al.,Calorimetry with deep learning: particle simulation and reconstruction for collider physics,Eur. Phys. J. C80(2020) 688 [1912.06794]

    Pith/arXiv arXiv 2020

  11. [11]

    Alimena, Y

    J. Alimena, Y. Iiyama and J. Kieseler,Fast convolutional neural networks for identifying long-lived particles in a high-granularity calorimeter,JINST15(2020) P12006 [2004.10744]

    Pith/arXiv arXiv 2020

  12. [12]

    Akchurin, C

    N. Akchurin, C. Cowden, J. Damgov, A. Hussain and S. Kunori,On the use of neural networks for energy reconstruction in high-granularity calorimeters,JINST16(2021) P12036 [2107.10207]

    Pith/arXiv arXiv 2021

  13. [13]

    Polson, L

    L. Polson, L. Kurchaninov and M. Lefebvre,Energy reconstruction in a liquid argon calorimeter cell using convolutional neural networks,JINST17(2022) P01002 [2109.05124]

    Pith/arXiv arXiv 2022

  14. [14]

    Khattak, S

    G.R. Khattak, S. Vallecorsa, F. Carminati and G.M. Khan,Fast simulation of a high granularity calorimeter by generative adversarial networks,Eur. Phys. J. C82(2022) 386 [2109.07388]

    Pith/arXiv arXiv 2022

  15. [15]

    Krause and D

    C. Krause and D. Shih,Fast and accurate simulations of calorimeter showers with normalizing flows,Phys. Rev. D107(2023) 113003 [2106.05285]

    Pith/arXiv arXiv 2023

  16. [16]

    Bieringer, A

    S. Bieringer, A. Butter, S. Diefenbacher, E. Eren, F. Gaede, D. Hundhausen et al., Calomplification — the power of generative calorimeter models,JINST17(2022) P09028 [2202.07352]

    Pith/arXiv arXiv 2022

  17. [17]

    Rogachev and F

    A. Rogachev and F. Ratnikov,GAN with an Auxiliary Regressor for the Fast Simulation of the Electromagnetic Calorimeter Response,J. Phys. Conf. Ser.2438(2023) 012086 [2207.06329]. [26]NA62collaboration,Improved calorimetric particle identification in NA62 using machine learning techniques,JHEP11(2023) 138 [2304.10580]

    Pith/arXiv arXiv 2023

  18. [18]

    S. Song, J. Chen, J. Liu, Y. Liu, B. Qi, Y. Shi et al.,Study of residual artificial neural network for particle identification in the CEPC high-granularity calorimeter prototype, JINST19(2024) P04033 [2310.09489]

    Pith/arXiv arXiv 2024

  19. [19]

    Charan,Particle identification with the Belle II calorimeter using machine learning,J

    A.N. Charan,Particle identification with the Belle II calorimeter using machine learning,J. Phys. Conf. Ser.2438(2023) 012111 [2301.11654]

    Pith/arXiv arXiv 2023

  20. [20]

    Simkina, F

    P. Simkina, F. Couderc, J. Malcl` es and M.¨O. Sahin,Reconstruction of electromagnetic showers in calorimeters using Deep Learning,Eur. Phys. J. C84(2024) 639 [2311.17914]

    Pith/arXiv arXiv 2024

  21. [21]

    Dubi´ nski, K

    J. Dubi´ nski, K. Deja, S. Wenzel, P. Rokita and T. Trzci´ nski,Machine learning methods for simulating particle response in the zero degree calorimeter at the ALICE experiment, CERN, AIP Conf. Proc.3061(2024) 040001 [2306.13606]

    Pith/arXiv arXiv 2024

  22. [22]

    Simsek, B

    E. Simsek, B. Isildak, A. Dogru, R. Aydogan, A.B. Bayrak and S. Ertekin,CALPAGAN: Calorimetry for Particles Using Generative Adversarial Networks,PTEP2024(2024) 083C01 [2401.02248]

    Pith/arXiv arXiv 2024

  23. [23]

    Hashmani, E

    R.K. Hashmani, E. Akba¸ s and M.B. Demirk¨ oz,A comparison of deep learning models for proton background rejection with the AMS electromagnetic calorimeter,Mach. Learn. Sci. Tech.5(2024) 045008 [2402.16285]

    Pith/arXiv arXiv 2024

  24. [24]

    Favaro, A

    L. Favaro, A. Ore, S.P. Schweitzer and T. Plehn,CaloDREAM – Detector Response Emulation via Attentive flow Matching,SciPost Phys.18(2025) 088 [2405.09629]

    Pith/arXiv arXiv 2025

  25. [25]

    Q. Liu, C. Shimmin, X. Liu, E. Shlizerman, S. Li and S.-C. Hsu,Calo-VQ: Vector-Quantized – 24 – Two-Stage Generative Model in Calorimeter Simulation,arXiv preprint(2024) [2405.06605]

    Pith/arXiv arXiv 2024

  26. [26]

    Qasim, N

    S.R. Qasim, N. Chernyavskaya, J. Kieseler, K. Long, O. Viazlo, M. Pierini et al.,End-to-end multi-particle reconstruction in high occupancy imaging calorimeters with graph neural networks,Eur. Phys. J. C82(2022) 753 [2204.01681]. [36]Belle-IIcollaboration,Photon Reconstruction in the Belle II Calorimeter Using Graph Neural Networks,Comput. Softw. Big Sci....

    Pith/arXiv arXiv 2022

  27. [27]

    Kobylianskii, N

    D. Kobylianskii, N. Soybelman, E. Dreyer and E. Gross,Graph-based diffusion model for fast shower generation in calorimeters with irregular geometry,Phys. Rev. D110(2024) 072003 [2402.11575]

    Pith/arXiv arXiv 2024

  28. [28]

    Akchurin et al.,Vertex Imaging Hadron Calorimetry Using AI/ML Tools,EPJ Web Conf

    N. Akchurin et al.,Vertex Imaging Hadron Calorimetry Using AI/ML Tools,EPJ Web Conf. 320(2025) 00026 [2408.15385]

    Pith/arXiv arXiv 2025

  29. [29]

    Acosta, B

    F.T. Acosta, B. Karki, P. Karande, A. Angerami, M. Arratia, K. Barish et al.,The optimal use of segmentation for sampling calorimeters,JINST19(2024) P06002 [2310.04442]

    Pith/arXiv arXiv 2024

  30. [30]

    Buhmann, S

    E. Buhmann, S. Diefenbacher, E. Eren, F. Gaede, G. Kasieczka, A. Korol et al.,CaloClouds: fast geometry-independent highly-granular calorimeter simulation,JINST18(2023) P11025 [2305.04847]

    Pith/arXiv arXiv 2023

  31. [31]

    Schnake, D

    S. Schnake, D. Kr¨ ucker and K. Borras,CaloPointFlow II Generating Calorimeter Showers as Point Clouds,arXiv preprint(2024) [2403.15782]

    Pith/arXiv arXiv 2024

  32. [32]

    Dimitrova, V

    K. Dimitrova, V. Kozhuharov, R. Nastaev and P. Petkov,Cluster Reconstruction in Electromagnetic Calorimeters Using Machine Learning Methods,J. Phys. Conf. Ser.3116 (2025) 012004 [2505.24740]

    Pith/arXiv arXiv 2025

  33. [33]

    Amram et al.,CaloChallenge 2022: a community challenge for fast calorimeter simulation,Rept

    O. Amram et al.,CaloChallenge 2022: a community challenge for fast calorimeter simulation,Rept. Prog. Phys.88(2025) 116201 [2410.21611]

    arXiv 2022

  34. [34]

    Everingham, L

    M. Everingham, L. Van Gool, C.K. Williams, J. Winn and A. Zisserman,The pascal visual object classes (voc) challenge,International journal of computer vision88(2010) 303

    2010

  35. [35]

    T.-Y. Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan et al.,Microsoft coco: Common objects in context, inEuropean conference on computer vision, pp. 740–755, Springer, 2014, DOI

    2014

  36. [36]

    Z. Wang, Y. Liu, Y. Tian, Y. Liu, Y. Wang and Q. Ye,Building vision models upon heat conduction, inProceedings of the Computer Vision and Pattern Recognition Conference, pp. 9707–9717, 2025, DOI

    2025

  37. [37]

    T.-Y. Lin, P. Goyal, R. Girshick, K. He and P. Doll´ ar,Focal loss for dense object detection, IEEE Transactions on Pattern Analysis and Machine Intelligence42(2018) 318

    2018

  38. [38]

    Vaswani, N

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A.N. Gomez et al.,Attention is all you need,Advances in neural information processing systems30(2017) . [50]GEANT4collaboration,GEANT4–a simulation toolkit,Nucl. Instrum. Meth. A506(2003) 250. – 25 –

    2017

  39. [39]

    L. Liu, X. Zhou and H. Peng,Development of a data-driven method to simulate the detector response of anti-neutron at besiii,Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment1033(2022) 166672

    2022

  40. [40]

    Gunderson, J

    B. Gunderson, J. Learned, J. Mapp and D.D. Reeder,A Measurement of the Anti-neutron - Proton Cross-section as Low-energy,Phys. Rev. D23(1981) 587. [53]BROOKHA VEN-HOUSTON-PENNSYL V ANIA STATE-RICEcollaboration, Measurement of Anti-neutron Proton Total and Annihilation Cross-sections From 100-MeV/c to 500-MeV/c,Phys. Rev. D36(1987) 659. [54]OBELIXcollabora...

    Pith/arXiv arXiv 1981

  41. [41]

    Cavallari,Performance of calorimeters at the lhc,Journal of Physics: Conference Series 293(2011) 012001

    F. Cavallari,Performance of calorimeters at the lhc,Journal of Physics: Conference Series 293(2011) 012001. [59]Particle Data Groupcollaboration,Review of particle physics,Phys. Rev. D110(2024) 030001

    2011

  42. [42]

    Sakharov,Violation of CP Invariance, C asymmetry, and baryon asymmetry of the universe,Pisma Zh

    A.D. Sakharov,Violation of CP Invariance, C asymmetry, and baryon asymmetry of the universe,Pisma Zh. Eksp. Teor. Fiz.5(1967) 32. [61]Bellecollaboration,The Belle Detector,Nucl. Instrum. Meth. A479(2002) 117. [62]KLOEcollaboration,Precision Kaon and Hadron Physics with KLOE,Riv. Nuovo Cim.31 (2008) 531 [0811.1929]. [63]CLAScollaboration,The CEBAF Large Ac...

    Pith/arXiv arXiv 1967

  43. [43]

    Anderle et al.,Electron-ion collider in China,Front

    D.P. Anderle et al.,Electron-ion collider in China,Front. Phys. (Beijing)16(2021) 64701 [2102.09222]. [67]CMScollaboration,The CMS Experiment at the CERN LHC,JINST3(2008) S08004. [68]ATLAScollaboration,The ATLAS Experiment at the CERN Large Hadron Collider,JINST 3(2008) S08003. [69]ALICEcollaboration,The ALICE experiment at the CERN LHC,JINST3(2008) S0800...

    Pith/arXiv arXiv 2021

  44. [44]

    Mixed-representation calorimetric network

    H. Yu et al., “Mixed-representation calorimetric network.” https://github.com/yuhongtian17/ViC, 2025

    2025

  45. [45]

    Z. Liu, Y. Lin, Y. Cao, H. Hu, Y. Wei, Z. Zhang et al.,Swin transformer: Hierarchical vision transformer using shifted windows, inProceedings of the IEEE/CVF international conference on computer vision, pp. 10012–10022, 2021, DOI

    2021

  46. [46]

    T.-Y. Lin, P. Doll´ ar, R. Girshick, K. He, B. Hariharan and S. Belongie,Feature pyramid networks for object detection, inProceedings of the IEEE conference on computer vision and pattern recognition, pp. 2117–2125, 2017, DOI

    2017

  47. [47]

    Devlin, M.-W

    J. Devlin, M.-W. Chang, K. Lee and K. Toutanova,Bert: Pre-training of deep bidirectional transformers for language understanding, inProceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologies, volume 1 (long and short papers), pp. 4171–4186, 2019, DOI

    2019

  48. [48]

    K. He, X. Chen, S. Xie, Y. Li, P. Doll´ ar and R. Girshick,Masked autoencoders are scalable vision learners, inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 16000–16009, 2022, DOI. – 27 –

    2022