pith. machine review for the scientific record. sign in

arxiv: 2604.22462 · v1 · submitted 2026-04-24 · ✦ hep-ph · astro-ph.CO· gr-qc

Recognition: unknown

Machine Learning for Multi-messenger Probes of New Physics and Cosmology: A Review and Perspective

Andrea Addazi , Konstantin Belotsky , Vitaly Beylin , Timur Bikbaev , Deen Chen , Filippo Fabrocini , Stefano Giagu , Krid Jinklub
show 13 more authors
Artem Kharakhashyan Maxim Khlopov Vladimir Korchagin Maxim Krasnov Atharv Mahajan Antonino Marciano Andrey Mayorov Antonio Morais Roman Pasechnik Jackson Levi Said Danila Sopin Viktor Stasenko Oem Trivedi
Authors on Pith no claims yet

Pith reviewed 2026-05-08 10:59 UTC · model grok-4.3

classification ✦ hep-ph astro-ph.COgr-qc
keywords machine learningmulti-messengerdark matternew physicscosmologygravitational wavescosmic raysneutrinos
0
0 comments X

The pith

Machine learning provides the key to integrating multi-messenger observations for probing dark matter and new physics.

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

The paper reviews how machine learning can enhance multi-messenger studies involving gravitational waves, cosmic rays, gamma rays, neutrinos, and collider experiments to investigate dark matter and physics beyond the Standard Model. It outlines a research program with three main objectives: analyzing new physics in cosmology including dark matter models, studying new physics signatures in cosmic ray experiments with cross-correlations, and developing machine learning methods for cosmic ray data analysis. The authors argue that this integrated approach will be essential for answering open questions in fundamental physics by combining information from different data sources. A sympathetic reader would care because it offers a concrete plan for future analyses that could reveal dark matter's properties and origins in ways not possible with isolated studies.

Core claim

The authors propose that machine learning can serve as the tool to unify heterogeneous multi-messenger datasets into a coherent inference framework, enabling extraction of information on dark matter properties, interactions, and genesis through combined analysis of gravitational waves, cosmic rays, gamma rays, neutrinos, and laboratory data.

What carries the argument

A unified inference framework powered by machine learning that integrates data from multiple messengers while preserving details from each.

If this is right

  • Multi-messenger analysis of new physics in cosmology, focusing on dark matter models.
  • Phenomenology of new physics signatures in ground-based cosmic ray experiments cross-correlated with other observations.
  • Development of machine learning techniques specifically for analyzing cosmic ray data in the context of new physics.
  • Improved constraints on dark matter from combining signals across gravitational waves, neutrinos, and colliders.

Where Pith is reading between the lines

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

  • Such a framework might allow predictions for signals in future detectors that combine multiple channels.
  • Extending this to include other cosmological probes like the cosmic microwave background could strengthen the inferences.
  • Validation on mock datasets with injected new physics signals would be a necessary first step before applying to real data.

Load-bearing premise

Machine learning methods can integrate data from different physical messengers into one framework that produces reliable new information on dark matter without adding biases or losing unique details from any single messenger.

What would settle it

If applying the proposed machine learning integration to known dark matter candidates or null results produces contradictory constraints compared to standard single-messenger analyses, that would indicate the approach fails to work as claimed.

Figures

Figures reproduced from arXiv: 2604.22462 by Andrea Addazi, Andrey Mayorov, Antonino Marciano, Antonio Morais, Artem Kharakhashyan, Atharv Mahajan, Danila Sopin, Deen Chen, Filippo Fabrocini, Jackson Levi Said, Konstantin Belotsky, Krid Jinklub, Maxim Khlopov, Maxim Krasnov, Oem Trivedi, Roman Pasechnik, Stefano Giagu, Timur Bikbaev, Viktor Stasenko, Vitaly Beylin, Vladimir Korchagin.

Figure 1
Figure 1. Figure 1: FIG. 1: Demonstration of likelihood-free inference (LFI) using a simulation-based approximate posterior in a 1D toy cosmologi view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2: Convolutional Neural Network (CNN) trained to estimate the cosmological parameter Ω view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3: Figure 6, from [204], shows a schematic representation of the PINN structure. This is relevant to tackle inverse problems view at source ↗
read the original abstract

The multi-messenger exploration of dark matter and physics beyond the Standard Model has emerged as a central direction in modern astro-particle physics, particularly following the discovery of gravitational waves. In this work, we present a comprehensive review and forward-looking perspective on machine-learning-enhanced multi-messenger approaches, combining information from gravitational waves, cosmic rays, gamma rays, neutrinos, and collider experiments. We summarize the current state of the field, discuss recent methodological developments, and outline a coherent research program aimed at integrating heterogeneous datasets within a unified inference framework. Our collaboration proposes here a plan for forthcoming analyses aiming at extracting information on the properties and interactions of dark matter, and finally on its genesis, combining multi-messenger astronomy techniques and inputs from laboratory physics. The main objectives planned in this line of research comprise: i) the multi-messenger analysis of new physics in cosmology, including mainly, but not only, several different models of dark matter; ii) the phenomenology of new physics signatures in ground-based cosmic rays experiments, with cross-correlation to the corresponding physical, astrophysical and cosmological observations; iii) the development of machine learning methods for data analysis in ground-based cosmic rays experiments, in light of the new physics signatures. We note that several groups have explored the use of multi-messenger observations, including gravitational waves, to probe alternative dark matter candidates. The present work builds on these developments by focusing on the role of machine learning in integrating heterogeneous datasets. We foresee that such a cross-fertilizing approach will represent the right path to extract information about the main questions left in fundamental physics.

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

0 major / 2 minor

Summary. The manuscript is a review and perspective paper summarizing the current state of machine-learning applications in multi-messenger probes of dark matter and physics beyond the Standard Model, incorporating data from gravitational waves, cosmic rays, gamma rays, neutrinos, and colliders. It outlines a proposed research program with three main objectives: multi-messenger analysis of new physics in cosmology (focusing on dark matter models), phenomenology of new physics signatures in ground-based cosmic ray experiments with cross-correlations, and development of ML methods for data analysis in those experiments, with the goal of integrating heterogeneous datasets in a unified inference framework to extract information on dark matter properties and genesis.

Significance. If the proposed ML-driven integration of multi-messenger datasets can be implemented without introducing biases or losing messenger-specific details, the perspective could provide a valuable roadmap for advancing astro-particle physics by enabling more powerful unified analyses that combine astronomical observations with laboratory inputs. The review component helps consolidate disparate literature on ML in this domain, and the forward-looking objectives highlight concrete directions for addressing open questions in fundamental physics.

minor comments (2)
  1. [Abstract] Abstract: The statement that 'several groups have explored the use of multi-messenger observations, including gravitational waves, to probe alternative dark matter candidates' would benefit from one or two specific citations to prior works in the main text to better ground the claim that the present work builds on these developments by focusing on ML integration.
  2. [Abstract] Abstract: The three listed objectives (i–iii) are logically structured, but the manuscript could clarify in the main text how these objectives will be pursued in a coordinated manner across the collaboration, e.g., by indicating shared datasets or sequential steps in the proposed analyses.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive and accurate summary of our review and perspective paper, as well as for recognizing the potential significance of the proposed ML-driven multi-messenger integration program. We appreciate the recommendation for minor revision and are pleased that the manuscript is viewed as a valuable consolidation of the literature with concrete forward-looking objectives.

Circularity Check

0 steps flagged

No significant circularity in this review paper

full rationale

This manuscript is a review and forward-looking perspective that summarizes external literature on multi-messenger observations of dark matter and new physics, then outlines a proposed research program for ML integration. No derivations, equations, fitted parameters, or predictions are presented anywhere in the text. The strongest claim is explicitly prospective ('we foresee that such a cross-fertilizing approach will represent the right path'), not an assertion that current results already succeed. All cited work is external; no self-citation chain is load-bearing for any internal result, and no ansatz or uniqueness theorem is smuggled in. The paper is therefore self-contained against external benchmarks with no reduction of outputs to its own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The paper rests on standard domain assumptions in multi-messenger astronomy and machine learning applicability to scientific data, without introducing new free parameters or entities.

axioms (1)
  • domain assumption Heterogeneous datasets from gravitational waves, cosmic rays, gamma rays, neutrinos, and colliders contain complementary information about dark matter that can be integrated via machine learning.
    This underpins the proposed unified inference framework and research plan.

pith-pipeline@v0.9.0 · 5684 in / 1226 out tokens · 34426 ms · 2026-05-08T10:59:13.374156+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

293 extracted references · 151 canonical work pages · 10 internal anchors

  1. [1]

    In this approach, the final state is represented as a probability density function (PDF) describing the likelihood of finding a nucleon at a given point in phase space

    Reliable models have been developed to simulate nuclear interactions in contexts such as ion therapy, including the Boltzmann–Langevin One Body (BLOB) model or efforts in QCD matter at extreme conditions [200], which describes heavy-ion interactions up to a few hundreds of MeV [211]. In this approach, the final state is represented as a probability densit...

  2. [2]

    A second major development is the deployment of machine learning algorithms at the trigger level, including im- plementations on FPGAs and heterogeneous hardware. At the LHC and future colliders, ML-enhanced triggers enable real-time event selection under severe latency and bandwidth constraints, allowing rare or unconven- tional signatures to be retained...

  3. [3]

    Another important class of applications concerns model-agnostic searches for new physics, where the goal is to detect deviations from the Standard Model without committing to specific signal hypotheses. Recently proposed approaches [214] based on anomaly-detection frameworks exploit unsupervised or weakly supervised learning to identify statistically sign...

  4. [4]

    A further application concerns the assessment of hadronic interaction models used in air-shower reconstruction from very-high-energy cosmic-ray observations, namely SIBYLL 2.3c [215], QGSJet II-04 [216], and EPOS-LHC [217]. A systematic scan of the corresponding output spaces can be performed using graph neural networks, which have already proven effectiv...

  5. [5]

    hidden valley

    Assessment of the three hadronic models hitherto applied to air-shower reconstructions from VHECR observa- tions, i.e. Sybil 2.3c [215], QGSJet II-04 [216] and EPOS-LHC [217], and consequently a scan of the output space, can be achieved within this perspective applying peep neural networks based on graph-neural network already used for classification prob...

  6. [6]

    A broad review of deep learning applications to LHC physics can be found in [198]

    A general and essential feature of machine-learning applications to collider physics is the ability to handle high-dimensional ensembles of data and to perform fast, real-time processing. A broad review of deep learning applications to LHC physics can be found in [198]. Techniques developed for collider experiments can therefore be directly transferred to...

  7. [7]

    gamma-ray spectra associated with dark matter annihilation or decay,

  8. [8]

    high-energy neutrino fluxes compatible with IceCube-like sensitivities,

  9. [9]

    dark photons

    stochastic gravitational-wave spectra generated by first-order phase transitions. Each observational channel is encoded through a dedicated feature extractor (e.g. convolutional or fully connected neural networks), producing latent representations that are subsequently combined within a joint inference architec- ture. The resulting model maps multi-messen...

  10. [10]

    The multi-messenger approach and cross-correlation with other channels The multi-messenger strategy requires to process jointly information arising from the gravitational radiation channel, the electromagnetic radiation channel, the cosmic rays and neutrino channel and the terrestrial collider channel. This strategy involves methods of cosmoparticle physi...

  11. [11]

    From the scientific community perspective, this trend started to emerge from the last decade, because of the rapid investment on many types of experiment/sections

    Interaction Networks, First Order Phase Transitions and BSM models We are currently living in the data/information era, with new trends emerging from the creation to storing to finally reading and interpreting the information conveyed by the data itself. From the scientific community perspective, this trend started to emerge from the last decade, because ...

  12. [12]

    extended

    in a Graph Network, to learn models of particle systems at different domains. In this respect, a similar approach can be deployed to tackle the rather challenging problem of using the combined information from particle collisions, cosmic rays and future space-based GWs detectors as a way to probe and constrain BSM theories. To have a better understanding ...

    2080

  13. [13]

    development of joint multi-messenger pipelines combining data from LHAASO, IceCube, Fermi-LAT and gravitational-wave observatories

  14. [14]

    implementation of machine learning frameworks for global parameter inference across heterogeneous datasets

  15. [15]

    systematic exploration of benchmark dark matter models within this unified framework

  16. [16]

    application of anomaly-detection techniques for model-independent searches

  17. [17]

    These directions outline a possible path toward transforming multi-messenger astrophysics

    investigation of next-generation methods, including physics-informed and quantum-enhanced machine learning. These directions outline a possible path toward transforming multi-messenger astrophysics. In conclusion of this review of theoretical and numerical methods, we mention a few urgent points to be developed in the multi-messenger investigation on DM. ...

    2026

  18. [18]

    Planck Collaboration, Astronomy & Astrophysics641, A1 (2020), 1807.06205

    work page arXiv 2020

  19. [19]

    ACT Collaboration, Journal of Cosmology and Astroparticle Physics12, 047 (2020), 2007.07288

    work page arXiv 2020

  20. [20]

    Li et al., National Science Review6, 145 (2019)

    H. Li et al., National Science Review6, 145 (2019)

    2019

  21. [21]

    Addazi, K

    A. Addazi, K. Belotsky, V. Beylin, et al., in37th International Cosmic Ray Conference (ICRC 2021)(2021), p. 474

    2021

  22. [22]

    HAWC Collaboration, Astrophysical Journal843, 39 (2017), 1701.01778

    work page arXiv 2017

  23. [23]

    Collaboration, Astronomy & Astrophysics457, 899 (2006), astro-ph/0607333

    H.E.S.S. Collaboration, Astronomy & Astrophysics457, 899 (2006), astro-ph/0607333

    work page arXiv 2006

  24. [24]

    M. G. Aartsen, M. Ackermann, et al., Physical Review Letters113, 101101 (2014), 1405.5303

    work page arXiv 2014

  25. [25]

    Evans and P

    L. Evans and P. Bryant, Journal of Instrumentation3, S08001 (2008)

    2008

  26. [26]

    Apollinari et al., CERN Yellow Reports: Monographs4(2017)

    G. Apollinari et al., CERN Yellow Reports: Monographs4(2017)

    2017

  27. [27]

    CEPC Study Group, arXiv:1811.10545 (2018)

    work page Pith review arXiv 2018

  28. [28]

    B. P. Abbott, R. Abbott, et al., Physical Review Letters116, 061102 (2016)

    2016

  29. [29]

    Laser Interferometer Space Antenna

    P. Amaro-Seoane et al.,Laser interferometer space antenna(2017), 1702.00786

    work page internal anchor Pith review arXiv 2017

  30. [30]

    Kawamura et al., Classical and Quantum Gravity28, 094011 (2011)

    S. Kawamura et al., Classical and Quantum Gravity28, 094011 (2011)

    2011

  31. [31]

    Crowder and N

    J. Crowder and N. J. Cornish, Physical Review D72, 083005 (2005)

    2005

  32. [32]

    Nan et al., International Journal of Modern Physics D20, 989 (2011)

    R. Nan et al., International Journal of Modern Physics D20, 989 (2011)

    2011

  33. [33]

    P. E. Dewdney et al., Proceedings of the IEEE97, 1482 (2009)

    2009

  34. [34]

    Hobbs et al., Classical and Quantum Gravity27, 084013 (2010)

    G. Hobbs et al., Classical and Quantum Gravity27, 084013 (2010)

    2010

  35. [35]

    Di Sciascio, Nuclear and Particle Physics Proceedings279-281, 166 (2016), 1602.07600

    G. Di Sciascio, Nuclear and Particle Physics Proceedings279-281, 166 (2016), 1602.07600

    work page arXiv 2016

  36. [36]

    X. Bai, B. Y. Bi, X. J. Bi, Z. Cao, et al.,The large high altitude air shower observatory (lhaaso) science white paper (2019), 1905.02773

    work page arXiv 2019

  37. [37]

    G. G. Madejski and M. Sikora, Annual Review of Astronomy and Astrophysics54, 725 (2016)

    2016

  38. [38]

    Ajello et al., Astrophysical Journal Supplement Series (2022), 2209.12070

    M. Ajello et al., Astrophysical Journal Supplement Series (2022), 2209.12070

    work page arXiv 2022

  39. [39]

    Addazi, S

    A. Addazi, S. Ketov, M. Khlopov, and A. Marciano, International Journal of Modern Physics D27, 1841011 (2018)

    2018

  40. [43]

    J. E. Kim, Physical Review Letters43, 103 (1979)

    1979

  41. [44]

    M. A. Shifman, A. I. Vainshtein, and V. I. Zakharov, Nuclear Physics B166, 493 (1980)

    1980

  42. [45]

    A. R. Zhitnitsky, Soviet Journal of Nuclear Physics31, 260 (1980)

    1980

  43. [46]

    Wilczek, Physical Review Letters49, 1549 (1982)

    F. Wilczek, Physical Review Letters49, 1549 (1982)

    1982

  44. [47]

    Z. G. Berezhiani and M. Y. Khlopov, Zeitschrift f¨ ur Physik C49, 73 (1991)

    1991

  45. [48]

    Chikashige, R

    Y. Chikashige, R. N. Mohapatra, and R. D. Peccei, Physics Letters B98, 265 (1981)

    1981

  46. [49]

    G. B. Gelmini and M. Roncadelli, Physics Letters B99, 411 (1981)

    1981

  47. [50]

    Berezhiani, European Physical Journal C76, 705 (2016), 1507.05478

    Z. Berezhiani, European Physical Journal C76, 705 (2016), 1507.05478

    work page arXiv 2016

  48. [51]

    Witten, Physics Letters B149, 351 (1984)

    E. Witten, Physics Letters B149, 351 (1984)

    1984

  49. [52]

    J. P. Conlon, Journal of High Energy Physics05, 078 (2006), hep-th/0602233

    work page Pith review arXiv 2006

  50. [53]

    Axions In String Theory

    P. Svrcek and E. Witten, Journal of High Energy Physics06, 051 (2006), hep-th/0605206

    work page Pith review arXiv 2006

  51. [54]

    R. D. Peccei and H. R. Quinn, Physical Review Letters38, 1440 (1977)

    1977

  52. [55]

    R. D. Peccei and H. R. Quinn, Physical Review D16, 1791 (1977)

    1977

  53. [56]

    Weinberg, Physical Review Letters40, 223 (1978)

    S. Weinberg, Physical Review Letters40, 223 (1978)

    1978

  54. [57]

    Wilczek, Physical Review Letters40, 279 (1978)

    F. Wilczek, Physical Review Letters40, 279 (1978)

    1978

  55. [58]

    D. A. Dicus et al., Physical Review D18, 1829 (1978)

    1978

  56. [59]

    Raffelt and L

    G. Raffelt and L. Stodolsky, Physical Review D37, 1237 (1988)

    1988

  57. [60]

    Andriamonje et al., Journal of Cosmology and Astroparticle Physics0704, 010 (2007)

    S. Andriamonje et al., Journal of Cosmology and Astroparticle Physics0704, 010 (2007)

    2007

  58. [61]

    De Angelis, M

    A. De Angelis, M. Roncadelli, and O. Mansutti, Physical Review D76, 121301 (2007)

    2007

  59. [62]

    De Angelis, O

    A. De Angelis, O. Mansutti, M. Persic, and M. Roncadelli, Monthly Notices of the Royal Astronomical Society394, L21 (2009)

    2009

  60. [63]

    Mirizzi and D

    A. Mirizzi and D. Montanino, Journal of Cosmology and Astroparticle Physics12, 004 (2009)

    2009

  61. [64]

    Simet, D

    M. Simet, D. Hooper, and P. D. Serpico, Physical Review D77, 063001 (2008)

    2008

  62. [65]

    Mirizzi, G

    A. Mirizzi, G. G. Raffelt, and P. D. Serpico, Lecture Notes in Physics741, 115 (2008)

    2008

  63. [66]

    Nikishov, Soviet Physics JETP14, 393 (1962)

    A. Nikishov, Soviet Physics JETP14, 393 (1962). 32

    1962

  64. [67]

    G. G. Fazio and F. W. Stecker, Nature226, 135 (1970)

    1970

  65. [68]

    Aharonian,Very High Energy Cosmic Gamma Radiation(World Scientific, Singapore, 2004)

    F. Aharonian,Very High Energy Cosmic Gamma Radiation(World Scientific, Singapore, 2004)

    2004

  66. [69]

    Dolgov and K

    A. Dolgov and K. Freese, Physical Review D51, 2693 (1995)

    1995

  67. [70]

    Dolgov, K

    A. Dolgov, K. Freese, R. Rangarajan, and M. Srednicki, Physical Review D56, 6155 (1997)

    1997

  68. [71]

    A. G. Cohen and D. B. Kaplan, Nuclear Physics B308, 913 (1988)

    1988

  69. [72]

    Y. B. Zeldovich and I. D. Novikov, Astronomicheskii Zhurnal10, 602 (1967)

    1967

  70. [73]

    S. W. Hawking, Monthly Notices of the Royal Astronomical Society152, 75 (1971)

    1971

  71. [74]

    B. J. Carr and S. W. Hawking, Monthly Notices of the Royal Astronomical Society168, 399 (1974)

    1974

  72. [75]

    S. W. Hawking, I. G. Moss, and J. M. Stewart, Physical Review D26, 2681 (1982)

    1982

  73. [76]

    Carr and F

    B. Carr and F. K¨ uhnel, Annual Review of Nuclear and Particle Science70, 355 (2020)

    2020

  74. [77]

    B. Carr, K. Kohri, Y. Sendouda, and J. Yokoyama, Reports on Progress in Physics84, 116902 (2021)

    2021

  75. [78]

    De Luca, V

    V. De Luca, V. Desjacques, G. Franciolini, and A. Riotto, Journal of Cosmology and Astroparticle Physics11, 028 (2020)

    2020

  76. [79]

    K. M. Belotsky, V. I. Dokuchaev, Y. N. Eroshenko, E. A. Esipova, M. Y. Khlopov, L. A. Khromykh, A. A. Kirillov, V. V. Nikulin, S. G. Rubin, and I. V. Svadkovsky, European Physical Journal C79, 246 (2019)

    2019

  77. [80]

    Berezin, V

    V. Berezin, V. Dokuchaev, Y. Eroshenko, and A. Smirnov, Universe6, 158 (2020)

    2020

  78. [81]

    Garc´ ıa-Bellido and S

    J. Garc´ ıa-Bellido and S. Clesse, Physics of the Dark Universe19, 144 (2018)

    2018

  79. [82]

    Calcino, J

    J. Calcino, J. Garc´ ıa-Bellido, and T. M. Davis, Monthly Notices of the Royal Astronomical Society479, 2889 (2018)

    2018

  80. [83]

    S. W. Hawking, Physics Letters B231, 237 (1989)

    1989

Showing first 80 references.