Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: McGrae-Menge, Madox C., Pierce, Jacob R., Fiuza, Frederico, Alves, E. Paulo
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2506.14048
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915812642652160
author McGrae-Menge, Madox C.
Pierce, Jacob R.
Fiuza, Frederico
Alves, E. Paulo
author_facet McGrae-Menge, Madox C.
Pierce, Jacob R.
Fiuza, Frederico
Alves, E. Paulo
contents Machine learning is offering powerful new 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.
format Preprint
id arxiv_https___arxiv_org_abs_2506_14048
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Embedding physical symmetries into machine-learned reduced plasma physics models via data augmentation
McGrae-Menge, Madox C.
Pierce, Jacob R.
Fiuza, Frederico
Alves, E. Paulo
Plasma Physics
Machine learning is offering powerful new 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.
title Embedding physical symmetries into machine-learned reduced plasma physics models via data augmentation
topic Plasma Physics
url https://arxiv.org/abs/2506.14048