- Physical Review Journals
All JournalsPhysics Magazine

Physical Review Research

Reuse & Permissions

It is not necessary to obtain permission to reuse this article or its components as it is available under the terms of the Creative Commons Attribution 4.0 International license. This license permits unrestricted use, distribution, and reproduction in any medium, provided attribution to the author(s) and the published article's title, journal citation, and DOI are maintained. Please note that some figures may have been included with permission from other third parties. It is your responsibility to obtain the proper permission from the rights holder directly for these figures.

Export citation

Export citation

Choose format for download:

Download Citation
  • Open Access

Embedding physical symmetries into machine-learned reduced plasma physics models via data augmentation

M. C. McGrae-Menge1,*, J. R. Pierce1, F. Fiuza2, and E. P. Alves1,3,†

  • *Contact author: madoxmm@physics.ucla.edu
  • †Contact author: epalves@physics.ucla.edu
PDF

Phys. Rev. Research 8, 013200 – Published 23 February, 2026

DOI: https://doi.org/10.1103/9fq1-vpqz

Abstract

Machine learning is offering powerful tools for the development and discovery of reduced models of nonlinear, multiscale plasma dynamics from the data of first-principles kinetic simulations. However, ensuring the physical consistency of such models requires embedding fundamental symmetries of plasma dynamics. In this work, we explore a symmetry-embedding strategy based on data augmentation, where symmetry-preserving transformations (e.g., Lorentz and Galilean boosts) are applied to simulation data. Using both sparse regression and neural networks, we show that models trained on symmetry-augmented data more accurately infer the plasma fluid equations and pressure tensor closures from fully kinetic particle-in-cell simulations of magnetic reconnection. We show that this approach suppresses spurious inertial-frame-dependent correlations between dynamical variables, improves data efficiency, and significantly outperforms models trained without symmetry-augmented data, as well as commonly used theoretical pressure closure models. Our results establish symmetry-based data augmentation as a broadly applicable method for incorporating physical structure into machine-learned reduced plasma models.

View figure in article

Physics Subject Headings (PhySH)

Article Text

References (85)

  1. J. Nuckolls, L. Wood, A. Thiessen, and G. Zimmerman, Laser compression of matter to super-high densities: Thermonuclear (CTR) applications, Nature (London) 239, 139 (1972).
  2. J. Ongena, R. Koch, R. Wolf, and H. Zohm, Magnetic-confinement fusion, Nat. Phys. 12, 398 (2016).
  3. L. Sironi and A. Spitkovsky, Relativistic reconnection: An efficient source of non-thermal particles, Astrophys. J. 783, L21 (2014).
  4. F. Guo, Y.-H. Liu, W. Daughton, and H. Li, Particle acceleration and plasma dynamics during magnetic reconnection in the magnetically-dominated regime, Astrophys. J. 806, 167 (2015).
  5. G. W. Hammett and F. W. Perkins, Fluid moment models for Landau damping with application to the ion-temperature-gradient instability, Phys. Rev. Lett. 64, 3019 (1990).
  6. O. Ohia, J. Egedal, V. S. Lukin, W. Daughton, and A. Le, Demonstration of anisotropic fluid closure capturing the kinetic structure of magnetic reconnection, Phys. Rev. Lett. 109, 115004 (2012).
  7. L. Comisso and L. Sironi, Particle acceleration in relativistic plasma turbulence, Phys. Rev. Lett. 121, 255101 (2018).
  8. E. P. Alves, J. Zrake, and F. Fiuza, Efficient nonthermal particle acceleration by the kink instability in relativistic jets, Phys. Rev. Lett. 121, 245101 (2018).
  9. B. Laperre, J. Amaya, S. Jamal, and G. Lapenta, Identification of high order closure terms from fully kinetic simulations using machine learning, Phys. Plasmas 29, 032706 (2022).
  10. Y. Qin, J. Ma, M. Jiang, C. Dong, H. Fu, L. Wang, W. Cheng, and Y. Jin, Data-driven modeling of Landau damping by physics-informed neural networks, Phys. Rev. Res. 5, 033079 (2023).
  11. Z. Huang, C. Dong, and L. Wang, Machine-learning heat flux closure for multi-moment fluid modeling of nonlinear Landau damping, Proc. Natl. Acad. Sci. USA 122, e2419073122 (2025).
  12. A. S. Joglekar and A. G. R. Thomas, Machine learning of hidden variables in multiscale fluid simulation, Mach. Learn.: Sci. Technol. 4, 035049 (2023).
  13. N. Barbour, W. Dorland, I. G. Abel, and M. Landreman, Machine-learning closure for Vlasov-Poisson dynamics in Fourier-Hermite space, J. Plasma Phys. 91, E140 (2025).
  14. D. D. Carvalho, D. R. Ferreira, and L. O. Silva, Learning the dynamics of a one-dimensional plasma model with graph neural networks, Mach. Learn.: Sci. Technol. 5, 025048 (2024).
  15. J. Bongard and H. Lipson, Automated reverse engineering of nonlinear dynamical systems, Proc. Natl. Acad. Sci. USA 104, 9943 (2007).
  16. M. Schmidt and H. Lipson, Distilling free-form natural laws from experimental data, Science 324, 81 (2009).
  17. S.-M. Udrescu and M. Tegmark, AI Feynman: A physics-inspired method for symbolic regression, Sci. Adv. 6, eaay2631 (2020).
  18. M. Cranmer, A. Sanchez Gonzalez, P. Battaglia, R. Xu, K. Cranmer, D. Spergel, and S. Ho, Discovering symbolic models from deep learning with inductive biases, in Advances in Neural Information Processing Systems, edited by H. Larochelle, M. Ranzato, R. Hadsell, M. Balcan, and H. Lin (Curran Associates, Inc., Red Hook, NY, USA, 2020), Vol. 33, pp. 17429–17442.
  19. W.-X. Wang, R. Yang, Y.-C. Lai, V. Kovanis, and C. Grebogi, Predicting catastrophes in nonlinear dynamical systems by compressive sensing, Phys. Rev. Lett. 106, 154101 (2011).
  20. S. L. Brunton, J. L. Proctor, and J. N. Kutz, Discovering governing equations from data by sparse identification of nonlinear dynamical systems, Proc. Natl. Acad. Sci. USA 113, 3932 (2016).
  21. S. H. Rudy, S. L. Brunton, J. L. Proctor, and J. N. Kutz, Data-driven discovery of partial differential equations, Sci. Adv. 3, e1602614 (2017).
  22. H. Schaeffer, Learning partial differential equations via data discovery and sparse optimization, Proc. R. Soc. A 473, 20160446 (2017).
  23. E. P. Alves and F. Fiuza, Data-driven discovery of reduced plasma physics models from fully kinetic simulations, Phys. Rev. Res. 4, 033192 (2022).
  24. J. Donaghy and K. Germaschewski, In search of a data-driven symbolic multi-fluid ten-moment model closure, J. Plasma Phys. 89, 895890105 (2023).
  25. E. R. Ingelsten, M. C. McGrae-Menge, E. P. Alves, and I. Pusztai, Data-driven discovery of a heat flux closure for electrostatic plasma phenomena, J. Plasma Phys. 91, E64 (2025).
  26. I. Abramovic, E. Alves, and M. Greenwald, Data-driven model discovery for plasma turbulence modelling, J. Plasma Phys. 88, 895880604 (2022).
  27. M. Dam, M. Brøns, J. Juul Rasmussen, V. Naulin, and J. S. Hesthaven, Sparse identification of a predator-prey system from simulation data of a convection model, Phys. Plasmas 24, 022310 (2017).
  28. A. A. Kaptanoglu, K. D. Morgan, C. J. Hansen, and S. L. Brunton, Physics-constrained, low-dimensional models for magnetohydrodynamics: First-principles and data-driven approaches, Phys. Rev. E 104, 015206 (2021).
  29. M. Golden, K. Satapathy, and D. Psaltis, Scalable discovery of fundamental physical laws: Learning magnetohydrodynamics from 3d turbulence data, arXiv:2501.04094.
  30. A. A. Kaptanoglu, T. Qian, F. Wechsung, and M. Landreman, Permanent-magnet optimization for stellarators as sparse regression, Phys. Rev. Appl. 18, 044006 (2022).
  31. A. A. Kaptanoglu, C. Hansen, J. D. Lore, M. Landreman, and S. L. Brunton, Sparse regression for plasma physics, Phys. Plasmas 30, 033906 (2023).
  32. J. D. Lore, S. De Pascuale, P. Laiu, B. Russo, J. S. Park, J. M. Park, S. L. Brunton, J. N. Kutz, and A. A. Kaptanoglu, Time-dependent SOLPS-ITER simulations of the tokamak plasma boundary for model predictive control using SINDy, Nucl. Fusion 63, 046015 (2023).
  33. B. Bayón-Buján and M. Merino, Data-driven sparse modeling of oscillations in plasma space propulsion, Mach. Learn.: Sci. Technol. 5, 035057 (2024).
  34. B. Sanderse, P. Stinis, R. Maulik, and S. E. Ahmed, Scientific machine learning for closure models in multiscale problems: A review, Found. Data Sci. 7, 298 (2025).
  35. A. A. Vlasov, On vibration properties of electron gas, J. Exp. Theor. Phys. 8, 291 (1938).
  36. A. A. Vlasov, The vibrational properties of an electron gas, Usp. Fiz. Nauk 93, 444 (1967).
  37. R. Hazeltine and F. Waelbroeck, The Framework of Plasma Physics (CRC Press, Boca Raton, FL, 1998).
  38. D. R. Gurevich, M. R. Golden, P. A. Reinbold, and R. O. Grigoriev, Learning fluid physics from highly turbulent data using sparse physics-informed discovery of empirical relations (SPIDER), J. Fluid Mech. 996, A25 (2024).
  39. M. M. Bronstein, J. Bruna, T. Cohen, and P. Veličković, Geometric deep learning: Grids, groups, graphs, geodesics, and gauges, arXiv:2104.13478.
  40. T. Cohen and M. Welling, Group equivariant convolutional networks, in Proceedings of The 33rd International Conference on Machine Learning, edited by M. F. Balcan and K. Q. Weinberger (PMLR, New York, 2016), Vol. 48, pp. 2990–2999.
  41. J. Spinner, V. Bresó, P. de Haan, T. Plehn, J. Thaler, and J. Brehmer, Lorentz-equivariant geometric algebra transformers for high-energy physics, Adv. Neural Info. Proc. Sys. 37, 22178 (2024).
  42. M. Zhdanov, D. Ruhe, M. Weiler, A. Lucic, J. Brandstetter, and P. Forré, Clifford-steerable convolutional neural networks, in Proceedings of the 41st International Conference on Machine Learning, edited by R. Salakhutdinov, Z. Kolter, K. Heller, A. Weller, N. Oliver, J. Scarlett, and F. Berkenkamp (Proceedings of Machine Learning Research, 2024), Vol. 235. pp. 61203–61228.
  43. M. Raissi, P. Perdikaris, and G. Karniadakis, Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, J. Comput. Phys. 378, 686 (2019).
  44. W. Hao, Symmetry-regularized neural ordinary differential equations, arXiv:2311.16628.
  45. S. Chen, E. Dobriban, and J. H. Lee, A group-theoretic framework for data augmentation, J. Machine Learning Res. 21, 1 (2020).
  46. A. Krizhevsky, I. Sutskever, and G. E. Hinton, ImageNet classification with deep convolutional neural networks, in Advances in Neural Information Processing Systems, edited by F. Pereira, C. Burges, L. Bottou, and K. Weinberger (Curran Associates, Inc., Red Hook, NY, USA, 2012), Vol. 25.
  47. D. C. Cireşan, U. Meier, L. M. Gambardella, and J. Schmidhuber, Deep, big, simple neural nets for handwritten digit recognition, Neural Comput. 22, 3207 (2010).
  48. S. Monaco, S. Barresi, and D. Apiletti, Lorentz-invariant augmentation for high-energy physics deep learning models, in Machine Learning and Principles and Practice of Knowledge Discovery in Databases, edited by R. Meo and F. Silvestri (Springer Nature, Cham, 2025), pp. 368–376.
  49. E. N. Parker, Sweet’s mechanism for merging magnetic fields in conducting fluids, J. Geophys. Res. 62, 509 (1957).
  50. J. Dungey, LXXVI. Conditions for the occurrence of electrical discharges in astrophysical systems, London Edinburgh Dublin Philos. Mag. J. Sci. 44, 725 (1953).
  51. P. A. Sweet, The neutral point theory of solar flares, in Electromagnetic Phenomena in Cosmical Physics, IAU Symposium, edited by B. Lehnert (Cambridge University Press, London, UK, 1958), Vol. 6, p. 123.
  52. H. E. Petschek, Magnetic field annihilation, in NASA Special Publication, edited by W. N. Hess (NASA Special Publication, Washington, DC, 1964), Vol. 50, p. 425.
  53. J. Birn, J. F. Drake, M. A. Shay, B. N. Rogers, R. E. Denton, M. Hesse, M. Kuznetsova, Z. W. Ma, A. Bhattacharjee, A. Otto, and P. L. Pritchett, Geospace environmental modeling (GEM) magnetic reconnection challenge, J. Geophys. Res.: Space Phys. 106, 3715 (2001).
  54. E. G. Zweibel and M. Yamada, Magnetic reconnection in astrophysical and laboratory plasmas, Annu. Rev. Astron. Astrophys. 47, 291 (2009).
  55. Y.-H. Liu, P. Cassak, X. Li, M. Hesse, S.-C. Lin, and K. Genestreti, First-principles theory of the rate of magnetic reconnection in magnetospheric and solar plasmas, Commun. Phys. 5, 97 (2022).
  56. P. A. Cassak, Y.-H. Liu, and M. Shay, A review of the 0.1 reconnection rate problem, J. Plasma Phys. 83, 715830501 (2017).
  57. S. Selvi, O. Porth, B. Ripperda, F. Bacchini, L. Sironi, and R. Keppens, Effective resistivity in relativistic collisionless reconnection, Astrophys. J. 950, 169 (2023).
  58. R. A. Fonseca, L. O. Silva, F. S. Tsung, V. K. Decyk, W. Lu, C. Ren, W. B. Mori, S. Deng, S. Lee, T. Katsouleas, and J. C. Adam, OSIRIS: A three-dimensional, fully relativistic particle in cell code for modeling plasma based accelerators, in Computational Science—ICCS 2002, edited by P. M. A. Sloot, A. G. Hoekstra, C. J. K. Tan, and J. J. Dongarra (Springer, Berlin, 2002), pp. 342–351.
  59. R. A. Fonseca, S. F. Martins, L. O. Silva, J. W. Tonge, F. S. Tsung, and W. B. Mori, One-to-one direct modeling of experiments and astrophysical scenarios: Pushing the envelope on kinetic plasma simulations, Plasma Phys. Controlled Fusion 50, 124034 (2008).
  60. E. G. Harris, On a plasma sheath separating regions of oppositely directed magnetic field, Nuovo Cimento 23, 115 (1962).
  61. M. C. McGrae-Menge, Embedding physical symmetries into machine-learned reduced plasma physics models via data augmentation: Data, Zenodo (2025), http://doi.org/10.5281/zenodo.15498576.
  62. J. P. Crutchfield and B. S. McNamara, Equations of motion from a data series, Complex Syst. 1, 417 (1987).
  63. H. Schaeffer and S. G. McCalla, Sparse model selection via integral terms, Phys. Rev. E 96, 023302 (2017).
  64. P. A. K. Reinbold, D. R. Gurevich, and R. O. Grigoriev, Using noisy or incomplete data to discover models of spatiotemporal dynamics, Phys. Rev. E 101, 010203(R) (2020).
  65. D. A. Messenger and D. M. Bortz, Weak SINDy for partial differential equations, J. Comput. Phys. 443, 110525 (2021).
  66. A. Adadi, A survey on data‐efficient algorithms in big data era, J. Big Data 8, 24 (2021).
  67. A. Le, J. Egedal, W. Daughton, W. Fox, and N. Katz, Equations of state for collisionless guide-field reconnection, Phys. Rev. Lett. 102, 085001 (2009).
  68. J. Ng, A. Hakim, A. Bhattacharjee, A. Stanier, and W. Daughton, Simulations of anti-parallel reconnection using a nonlocal heat flux closure, Phys. Plasmas 24, 082112 (2017).
  69. J. Ng, A. Hakim, L. Wang, and A. Bhattacharjee, An improved ten-moment closure for reconnection and instabilities, Phys. Plasmas 27, 082106 (2020).
  70. G. F. Chew, M. L. Goldberger, F. E. Low, and S. Chandrasekhar, The Boltzmann equation and the one-fluid hydromagnetic equations in the absence of particle collisions, Proc. R. Soc. London A 236, 112 (1956).
  71. K. Hornik, M. Stinchcombe, and H. White, Multilayer feedforward networks are universal approximators, Neural Networks 2, 359 (1989).
  72. A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Köpf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala, PyTorch: an imperative style, high-performance deep learning library, in Proceedings of the 33rd International Conference on Neural Information Processing Systems (Curran Associates Inc., Red Hook, NY, USA, Article, 2019), Vol. 721, pp. 8026–8037.
  73. E. D. Held, J. D. Callen, C. C. Hegna, C. R. Sovinec, T. A. Gianakon, and S. E. Kruger, Nonlocal closures for plasma fluid simulations, Phys. Plasmas 11, 2419 (2004).
  74. R. Maulik, N. A. Garland, J. W. Burby, X.-Z. Tang, and P. Balaprakash, Neural network representability of fully ionized plasma fluid model closures, Phys. Plasmas 27, 072106 (2020).
  75. C. Ma, B. Zhu, X.-Q. Xu, and W. Wang, Machine learning surrogate models for Landau fluid closure, Phys. Plasmas 27, 042502 (2020).
  76. L. Bois, E. Franck, L. Navoret, and V. Vigon, A neural network closure for the Euler-Poisson system based on kinetic simulations, Kinet. Relat. Models 15, 49 (2022).
  77. M. C. McGrae-Menge, D. D. Carvalho, A. Velberg, M. Almanza, N. Barbour, V. K. Decyk, F. Fiuza, W. D. Dorland, N. F. Loureiro, and E. P. Alves, Data-driven discovery of fluid closures for magnetic reconnection from fully kinetic particle-in-cell simulations, in 66th Annual Meeting of the APS Division of Plasma Physics (American Physical Society, Atlanta, GA, 2024).
  78. J. Huang, Y. Cheng, A. J. Christlieb, and L. F. Roberts, Machine learning moment closure models for the radiative transfer equation I: Directly learning a gradient based closure, J. Comput. Phys. 453, 110941 (2022).
  79. R. Greif, F. Jenko, and N. Thuerey, Physics-preserving AI-accelerated simulations of plasma turbulence, arXiv:2309.16400.
  80. C. Chang, V. Deringer, K. Katti, V. Speybroeck, and C. Wolverton, Simulations in the era of exascale computing, Nat. Rev. Mater. 8, 309 (2023).
  81. M. Bugli, E. F. Lopresti, E. Figueiredo, A. Mignone, B. Cerutti, G. Mattia, L. Del Zanna, G. Bodo, and V. Berta, Relativistic reconnection with effective resistivity: I. Dynamics and reconnection rate, Astron. Astrophys. 693, A233 (2025).
  82. A. Moran, L. Sironi, A. Levis, B. Ripperda, E. R. Most, and S. Selvi, Effective resistivity in relativistic reconnection: A prescription based on fully kinetic simulations, Astrophys. J. Lett. 978, L45 (2025).
  83. K. Kaheman, J. N. Kutz, and S. L. Brunton, SINDy-PI: A robust algorithm for parallel implicit sparse identification of nonlinear dynamics, Proc. R. Soc. A 476, 20200279 (2020).
  84. M. Golden, Scalable sparse regression for model discovery: The fast lane to insight, arXiv:2405.09579.
  85. J.-L. Vay, D. L. Bruhwiler, C. G. R. Geddes, W. M. Fawley, S. F. Martins, J. R. Cary, E. Cormier-Michel, B. Cowan, R. A. Fonseca, M. A. Furman, W. Lu, W. B. Mori, and L. O. Silva, Simulating relativistic beam and plasma systems using an optimal boosted frame, J. Phys.: Conf. Ser. 180, 012006 (2009).

Outline

Information

Phys. Rev. Research 8, 013200– Published 23 February, 2026
  • Received 6 June 2025
  • Accepted 23 January 2026
Crossmark - Check for updates button
Creative Commons Logo

Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article's title, journal citation, and DOI.

Published by the American Physical Society

Sign In to Your Journals Account

Filter

Filter

Article Lookup

Enter a citation