EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

How Unconstrained Machine-Learning Models Learn Physical Symmetries

M. Domina, J. W. Abbott, P. Pegolo, F. Bigi, M. Ceriotti

PRX Intelligence · 2026

DOI: 10.1103/cnsd-p8nc ↗

Kurz gefasst

Fachartikel in der Zeitschrift PRX Intelligence, 2026. Verfasst von M. Domina und 4 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 61 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung, Centro Svizzero di Calcolo Scientifico.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The requirement of generating predictions that exactly fulfill the fundamental symmetry of the corresponding physical quantities has profoundly shaped the development of machine-learning (ML) models for physical simulations. In many cases, models are built using constrained mathematical forms that ensure that symmetries are enforced exactly. However, unconstrained models that do not obey rotational symmetries are often found to have competitive performance, and to be able to to a high level of accuracy an approximate equivariant behavior with a simple data augmentation strategy. In this paper, we introduce rigorous metrics to measure the symmetry content of the learned representations in such models, and assess the accuracy with which the outputs fulfill the equivariant condition. We apply these metrics to two unconstrained, transformer-based models operating on decorated point clouds (a graph neural network for atomistic simulations and a PointNet-style architecture for particle physics) to investigate how symmetry information is processed across architectural layers and is learned during training. Based on these insights, we establish a rigorous framework for diagnosing spectral failure modes in ML models. Enabled by this analysis, we demonstrate that one can achieve superior stability and accuracy by strategically injecting the minimum required inductive biases, preserving the high expressivity and scalability of unconstrained architectures while guaranteeing physical fidelity.

Förderung

  • Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung
  • Centro Svizzero di Calcolo Scientifico

Zitieren

APA

Domina, M., Abbott, J. W., Pegolo, P., Bigi, F., & Ceriotti, M. (2026). How Unconstrained Machine-Learning Models Learn Physical Symmetries. PRX Intelligence. https://doi.org/10.1103/cnsd-p8nc

Vancouver

Domina M, Abbott JW, Pegolo P, Bigi F, Ceriotti M. How Unconstrained Machine-Learning Models Learn Physical Symmetries. PRX Intelligence. 2026. doi:10.1103/cnsd-p8nc

BibTeX
@article{Domina2026,
  author = {Domina, M. and Abbott, J. W. and Pegolo, P. and Bigi, F. and Ceriotti, M.},
  title = {How Unconstrained Machine-Learning Models Learn Physical Symmetries},
  journal = {PRX Intelligence},
  year = {2026},
  doi = {10.1103/cnsd-p8nc}
}
Kostenlos starten →