pith. sign in

arxiv: 2606.23353 · v1 · pith:LQTT67UZnew · submitted 2026-06-22 · ⚛️ nucl-th · cs.LG· nucl-ex· physics.optics· quant-ph

Ultra-Peripheral Collisions as a Nuclear-Structure Interferometer with Interpretable Multitask Deep Learning

Jing-Zong Zhang , Wang-Mei Zha , Lingxiao Wang , Guo-Liang Ma This is my paper

Pith reviewed 2026-06-26 06:31 UTC · model grok-4.3

classification ⚛️ nucl-th cs.LGnucl-exphysics.opticsquant-ph
keywords ultra-peripheral collisionsnuclear structuredeep learningJ/ψ photoproductionnuclear deformationneutron skincoherent productionzirconium collisions
0
0 comments X

The pith

An interpretable multitask deep learning model extracts multiple nuclear structure indicators from transverse momentum distributions in ultra-peripheral collisions.

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

The paper establishes that ultra-peripheral collisions act as a nuclear-structure interferometer through coherent vector-meson photoproduction, where diffraction and two-source interference patterns directly encode the nuclear spatial density. It introduces an interpretable multitask deep-learning framework that simultaneously maps these distributions to several nuclear indicators while identifying the kinematic regions responsible for each output. Demonstrated on coherent J/ψ production in zirconium-96 collisions, the approach addresses the inverse problem posed by correlated sensitivities to deformation and neutron skin, phase smearing, and backgrounds. A sympathetic reader would care because precise nuclear structure data remain essential yet difficult to obtain across fundamental physics, and this method offers a route to quantitative constraints from high-luminosity data.

Core claim

The central claim is that an interpretable multitask deep-learning framework maps transverse momentum distributions from coherent J/ψ photoproduction in ultra-peripheral collisions to multiple nuclear-structure indicators at once, with the learned features separating diffraction-dominated from interference-dominated information and yielding analysis-ready observables.

What carries the argument

The interpretable multitask deep-learning framework that maps transverse momentum distributions to nuclear-structure indicators while identifying driving kinematic regions.

If this is right

  • The learned features separate diffraction-dominated and interference-dominated information in the transverse momentum distributions.
  • The framework supplies analysis-ready observables suitable for future high-luminosity data sets.
  • Multiple nuclear-structure indicators can be constrained simultaneously from the same photoproduction patterns.
  • Kinematic regions responsible for each indicator inference become identifiable without manual fitting.

Where Pith is reading between the lines

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

  • The same multitask architecture could be retrained on other vector mesons or nuclear species to test whether separation of diffraction and interference persists across different mass ranges.
  • If the model succeeds on real data, traditional one-at-a-time fitting procedures for UPC observables might be replaced by direct extraction of multiple parameters from a single network pass.
  • The identified kinematic regions could guide future detector design or trigger strategies to emphasize the interference-dominated regime.

Load-bearing premise

The simulated training data accurately capture all relevant experimental effects including phase smearing, correlated sensitivities to deformation and neutron skin, and backgrounds.

What would settle it

Applying the trained model to real high-luminosity UPC data and finding that the extracted deformation and neutron-skin values disagree with independent measurements from electron scattering or other probes would falsify the claim that the framework generalizes to produce reliable observables.

Figures

Figures reproduced from arXiv: 2606.23353 by Guo-Liang Ma, Jing-Zong Zhang, Lingxiao Wang, Wang-Mei Zha.

Figure 1
Figure 1. Figure 1: FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: displays the simulated J/ψ photoproduction probability in the transverse momentum plane at mid-rapidity (y = 0) with b = 12 fm. Focusing on the coherent process in [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3: Schematic of the Multitask deep-learning architecture. The pipeline accepts the two-dimensional transverse momentum [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6 [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7 [PITH_FULL_IMAGE:figures/full_fig_p008_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8 [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9 [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: FIG. 10 [PITH_FULL_IMAGE:figures/full_fig_p013_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: FIG. 11 [PITH_FULL_IMAGE:figures/full_fig_p014_11.png] view at source ↗
read the original abstract

Precise knowledge of nuclear structure is essential across fundamental physics, yet probing these structures is notoriously difficult. To address this challenge, ultra-peripheral collisions (UPCs) provide a femtoscopic tomography for imaging the atomic nucleus. UPCs offer a pristine electromagnetic pathway: coherent vector-meson photoproduction generates patterns of diffraction and two-source interference that directly encode the nuclear spatial density. Turning these patterns into quantitative constraints is, however, a challenging inverse problem, complicated by correlated sensitivities to deformation and neutron skin, phase smearing, and experimental backgrounds. Here we introduce an interpretable Multitask deep-learning framework that maps transverse momentum distributions to multiple nuclear-structure indicators simultaneously and identifies the kinematic regions driving each inference. We demonstrate the approach with coherent $J/\psi$ photoproduction in $^{96}_{40}\text{Zr} + ^{96}_{40}\text{Zr}$ collisions, showing that the learned features separate diffraction-dominated and interference-dominated information and provide analysis-ready observables for future high-luminosity data.

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

1 major / 0 minor

Summary. The manuscript introduces an interpretable multitask deep-learning framework that maps transverse momentum distributions from ultra-peripheral collisions to multiple nuclear-structure indicators simultaneously while identifying the driving kinematic regions. It demonstrates the method on simulated coherent J/ψ photoproduction in ^{96}Zr + ^{96}Zr collisions, claiming that the learned features separate diffraction-dominated and interference-dominated information and yield analysis-ready observables.

Significance. If the forward model in simulation faithfully encodes nuclear densities, phase information, and backgrounds, the multitask approach with built-in interpretability offers a practical route to handle correlated sensitivities to deformation and neutron skin in UPC data. The explicit separation of diffraction and interference channels is a concrete strength that could support future high-luminosity analyses.

major comments (1)
  1. The central claim that the learned features provide analysis-ready observables for real high-luminosity data rests on the untested assumption that the simulated training data captures all relevant experimental effects (phase smearing, correlated sensitivities, backgrounds). No section demonstrates robustness tests, held-out validation metrics, or sensitivity to variations in these inputs, which is load-bearing for generalization.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive and detailed report. The major comment identifies a key limitation in the current scope of validation, which we address directly below and will resolve through manuscript revision.

read point-by-point responses

  1. Referee: [—] The central claim that the learned features provide analysis-ready observables for real high-luminosity data rests on the untested assumption that the simulated training data captures all relevant experimental effects (phase smearing, correlated sensitivities, backgrounds). No section demonstrates robustness tests, held-out validation metrics, or sensitivity to variations in these inputs, which is load-bearing for generalization.

    Authors: We agree that the manuscript's forward-looking claim regarding analysis-ready observables would be strengthened by explicit robustness demonstrations. The current work is a proof-of-concept demonstration on forward-model simulations that encode nuclear densities, diffraction, and interference; it does not yet incorporate full experimental variations. In the revised version we will add a dedicated subsection reporting held-out validation metrics on parameter-varied simulations, sensitivity scans to phase smearing and background levels, and quantitative assessment of how these affect the extracted nuclear-structure indicators. This will clarify the present scope while supporting the intended application to future data. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper introduces a multitask deep-learning framework that maps simulated transverse-momentum distributions from coherent J/ψ photoproduction in Zr+Zr UPCs to multiple nuclear-structure indicators. The demonstration relies on the forward simulation faithfully encoding nuclear densities, interference phases, and backgrounds so that learned features remain informative on real data. No equations or steps in the provided abstract and description reduce a claimed prediction to a fitted input by construction, invoke a self-citation as the sole justification for a uniqueness claim, or rename a known empirical pattern as a new derivation. The central claim is an inverse-problem application whose validity rests on external simulation fidelity rather than internal self-reference; the derivation chain is therefore self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review; no explicit free parameters, axioms, or invented entities are stated in the provided text.

pith-pipeline@v0.9.1-grok · 5729 in / 995 out tokens · 26150 ms · 2026-06-26T06:31:08.411985+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

64 extracted references · 2 linked inside Pith

  1. [1]

    Conversely, a halo-type structure is characterized by a larger surface diffuseness for neutrons (a n > a p) with similar radii

    while keeping the diffusenessaidentical. Conversely, a halo-type structure is characterized by a larger surface diffuseness for neutrons (a n > a p) with similar radii. The coherent photoproduction amplitude is constructed by convolving the equivalent photon flux [38] with the photon-nucleus scattering amplitude, incorporating nuclear shadowing and cohere...

  2. [2]

    N. P. Nazirkar, X. Shi, J. Shi, M. N’Gom, and E. Fohtung, Appl. Phys. Rev.11, 021302 (2024)

    2024

  3. [3]

    Ossig, C

    C. Ossig, C. Strelow, J. Fl¨ ugge, S. Patjens, J. Garrevoet, K. Spiers, J. L. Barp, J. Hagemann, F. Seiboth, M. De Bastiani, et al., PRX Energy3, 023011 (2024)

    2024

  4. [4]

    J. Miao, P. Charalambous, J. Kirz, and D. Sayre, Nature400, 342 (1999)

    1999

  5. [5]

    Thibault, M

    P. Thibault, M. Dierolf, A. Menzel, O. Bunk, C. David, and F. Pfeiffer, Science321, 379 (2008)

    2008

  6. [6]

    Ophus, Microsc

    C. Ophus, Microsc. Microanal.25, 563 (2019)

    2019

  7. [7]

    Ozdol, C

    V. Ozdol, C. Gammer, X. Jin, P. Ercius, C. Ophus, J. Ciston, and A. Minor, Appl. Phys. Lett.106, 253107 (2015)

    2015

  8. [8]

    A. Yau, W. Cha, M. W. Kanan, G. B. Stephenson, and A. Ulvestad, Science356, 739 (2017)

    2017

  9. [9]

    Robinson and R

    I. Robinson and R. Harder, Nature Mater.8, 291 (2009)

    2009

  10. [10]

    C.-C. Chen, C. Zhu, E. R. White, C.-Y. Chiu, M. Scott, B. Regan, L. D. Marks, Y. Huang, and J. Miao, Nature496, 74 (2013)

    2013

  11. [11]

    J. A. Rodriguez, R. Xu, C.-C. Chen, Z. Huang, H. Jiang, A. L. Chen, K. S. Raines, A. Pryor Jr, D. Nam, L. Wiegart, et al., IUCrJ2, 575 (2015)

    2015

  12. [12]

    Hamamoto and B

    I. Hamamoto and B. R. Mottelson, Scholarpedia7, 10693 (2011)

    2011

  13. [13]

    Cline, Ann

    D. Cline, Ann. Rev. Nucl. Part. Sci.36, 683 (1986)

    1986

  14. [14]

    Ding, D.-Q

    M.-Q. Ding, D.-Q. Fang, and Y.-G. Ma, Nucl. Sci. Tech.35, 211 (2024)

    2024

  15. [15]

    F. J. Fattoyev and J. Piekarewicz, Phys. Rev. C86, 015802 (2012)

    2012

  16. [16]

    J. M. Lattimer and M. Prakash, Science304, 536 (2004)

    2004

  17. [17]

    Roca-Maza, M

    X. Roca-Maza, M. Centelles, X. Vinas, and M. Warda, Phys. Rev. Lett.106, 252501 (2011)

    2011

  18. [18]

    J. D. Patterson and R. J. Peterson, Nucl. Phys. A717, 235 (2003)

    2003

  19. [19]

    M. I. Abdulhamid et al. (STAR), Nature635, 67 (2024)

    2024

  20. [20]

    Ragoni (ALICE), Acta Phys

    S. Ragoni (ALICE), Acta Phys. Polon. Supp.18, 1 (2025)

    2025

  21. [21]

    K. Shen, X. Wu, Z. Tang, and W. Zha, Eur. Phys. J. C84, 1167 (2024)

    2024

  22. [22]

    J. Luo, X. Li, Z. Tang, X. Wu, and W. Zha, Phys. Rev. C108, 054906 (2023)

    2023

  23. [23]

    W. Zha, L. Ruan, Z. Tang, Z. Xu, and S. Yang, Phys. Rev. C99, 061901 (2019)

    2019

  24. [24]

    Lin, J.-Y

    S. Lin, J.-Y. Hu, H.-J. Xu, S. Pu, and Q. Wang, Phys. Rev. D111, 074020 (2025)

    2025

  25. [25]

    B. I. Abelev et al. (STAR), Phys. Rev. Lett.102, 112301 (2009)

    2009

  26. [26]

    Ma, Nucl

    Y.-G. Ma, Nucl. Sci. Tech.34, 16 (2023)

    2023

  27. [27]

    Aarts, K

    G. Aarts, K. Fukushima, T. Hatsuda, A. Ipp, S. Shi, L. Wang, and K. Zhou, Nature Rev. Phys.7, 154 (2025)

    2025

  28. [28]

    M. S. Abdallah et al. (STAR), Phys. Rev. Lett.128, 092301 (2022)

    2022

  29. [29]

    H.-J. Xu, H. Li, X. Wang, C. Shen, and F. Wang, Phys. Lett. B819, 136453 (2021)

    2021

  30. [30]

    J. Jia, G. Giacalone, and C. Zhang, Phys. Rev. Lett.131, 022301 (2023)

    2023

  31. [31]

    Boehnlein et al., Rev

    A. Boehnlein et al., Rev. Mod. Phys.94, 031003 (2022). 11

    2022

  32. [32]

    He, Y.-G

    W.-B. He, Y.-G. Ma, L.-G. Pang, H.-C. Song, and K. Zhou, Nucl. Sci. Tech.34, 88 (2023)

    2023

  33. [33]

    K. Zhou, L. Wang, L.-G. Pang, and S. Shi, Prog. Part. Nucl. Phys.135, 104084 (2024)

    2024

  34. [34]

    Ma, L.-G

    Y.-G. Ma, L.-G. Pang, R. Wang, and K. Zhou, Chin. Phys. Lett.40, 122101 (2023)

    2023

  35. [35]

    Y.-L. Du, D. Pablos, and K. Tywoniuk, Phys. Rev. Lett.128, 012301 (2022)

    2022

  36. [36]

    A. J. Larkoski, Eur. Phys. J. C84, 1117 (2024)

    2024

  37. [37]

    Utama, J

    R. Utama, J. Piekarewicz, and H. B. Prosper, Phys. Rev. C93, 014311 (2016)

    2016

  38. [38]

    Cheng, S

    Y.-L. Cheng, S. Shi, Y.-G. Ma, H. St¨ ocker, and K. Zhou, Phys. Rev. C107, 064909 (2023)

    2023

  39. [39]

    Krauss, M

    F. Krauss, M. Greiner, and G. Soff, Prog. Part. Nucl. Phys.39, 503 (1997)

    1997

  40. [40]

    M. L. Miller, K. Reygers, S. J. Sanders, and P. Steinberg, Ann. Rev. Nucl. Part. Sci.57, 205 (2007)

    2007

  41. [41]

    T. H. Bauer, R. D. Spital, D. R. Yennie, and F. M. Pipkin, Rev. Mod. Phys.50, 261 (1978)

    1978

  42. [42]

    P. Wang, X. Wu, W. Zha, and Z. Tang, Chin. Phys. C46, 074103 (2022)

    2022

  43. [43]

    S. R. Klein, J. Nystrand, J. Seger, Y. Gorbunov, and J. Butterworth, Comput. Phys. Commun.212, 258 (2017)

    2017

  44. [44]

    S. R. Klein and J. Nystrand, Phys. Rev. Lett.84, 2330 (2000)

    2000

  45. [45]

    K. He, X. Zhang, S. Ren, and J. Sun (2015), 1512.03385

    Pith/arXiv arXiv 2015

  46. [46]

    L. M. Zintgraf, T. S. Cohen, T. Adel, and M. Welling (2017), 1702.04595

    Pith/arXiv arXiv 2017

  47. [47]

    Grund (ALICE), Phys

    D. Grund (ALICE), Phys. Proc. UPC1, 4 (2024)

    2024

  48. [48]

    Massacrier (ALICE), in58th Rencontres de Moriond on QCD and High Energy Interactions(2024), 2407.09707

    L. Massacrier (ALICE), in58th Rencontres de Moriond on QCD and High Energy Interactions(2024), 2407.09707

    arXiv 2024

  49. [49]

    M. I. Abdulhamid et al. (STAR), Phys. Rev. Lett.133, 052301 (2024)

    2024

  50. [50]

    Accardi et al., Eur

    A. Accardi et al., Eur. Phys. J. A52, 268 (2016)

    2016

  51. [51]

    J. D. Brandenburg, S. R. Klein, Z. Xu, S. Yang, W. Zha, and J. Zhou, Prog. Part. Nucl. Phys.143, 104174 (2025)

    2025

  52. [52]

    Guzey, E

    V. Guzey, E. Kryshen, M. Strikman, and M. Zhalov, Phys. Lett. B726, 290 (2013)

    2013

  53. [53]

    Garcon, Eur

    M. Garcon, Eur. Phys. J. A18, 389 (2003)

    2003

  54. [54]

    M¨ antysaari and B

    H. M¨ antysaari and B. Schenke, Phys. Rev. Lett.117, 052301 (2016)

    2016

  55. [55]

    D. Y. Shao, H.-Q. Yu, C. Zhang, and J. Zhou (2025), 2511.17670

    arXiv 2025

  56. [56]

    Acharya et al

    S. Acharya et al. (ALICE), Phys. Lett. B865, 139466 (2025)

    2025

  57. [57]

    Shatat (ALICE), EPJ Web Conf.296, 07005 (2024)

    A. Shatat (ALICE), EPJ Web Conf.296, 07005 (2024)

    2024

  58. [58]

    Pavone, A

    A. Pavone, A. Merlo, S. Kwak, and J. Svensson, Plasma Phys. Control. Fusion65, 053001 (2023)

    2023

  59. [59]

    J. E. Bernhard, J. S. Moreland, S. A. Bass, J. Liu, and U. Heinz, Phys. Rev. C94, 024907 (2016)

    2016

  60. [60]

    He, W.-B

    J. He, W.-B. He, Y.-G. Ma, and S. Zhang, Phys. Rev. C104, 044902 (2021)

    2021

  61. [61]

    Pratt, E

    S. Pratt, E. Sangaline, P. Sorensen, and H. Wang, Phys. Rev. Lett.114, 202301 (2015)

    2015

  62. [62]

    Carleo, I

    G. Carleo, I. Cirac, K. Cranmer, L. Daudet, M. Schuld, N. Tishby, L. Vogt-Maranto, and L. Zdeborov´ a, Rev. Mod. Phys. 91, 045002 (2019)

    2019

  63. [63]

    Hufner and B

    J. Hufner and B. Z. Kopeliovich, Phys. Lett. B426, 154 (1998). 12 Appendix A: GLAUBER-VMD FRAMEWORK DET AIL J/ψphotoproduction in 96 40Zr+96 40Zr ultra-peripheral collisions is simulated using the Glauber-VMD framework established in Refs. [20–22]. The spatial distribution of nucleons inside the 96 40Zr nucleus is modeled with a deformed Woods–Saxon profi...

    1998

  64. [64]

    Conversely, a halo-type structure is characterized by a larger surface diffuseness for neutrons (a n > a p) with similar radii

    while keeping the diffusenessaidentical. Conversely, a halo-type structure is characterized by a larger surface diffuseness for neutrons (a n > a p) with similar radii. The coherent production amplitude is formulated by combining the equivalent photon flux with the nuclear scat- tering amplitude. The spatial photon flux induced by the projectile ion is gi...