Tangent-Plane Evidential Uncertainty in Active Learning for Magnetic Interatomic Potentials
Abstract
Magnetic interatomic potentials need to account for coupled lattice and spin degrees of freedom, yet constructing reliable training sets remains costly because noncollinear first-principles labels are expensive. Active learning can mitigate this cost, provided that the uncertainty estimate is physically meaningful for the magnetic-response targets that drive spin reorientation. Here we extend the e2IP evidential framework to magnetic machine-learning interatomic potentials by formulating the projected spin-force likelihood and the corresponding epistemic uncertainty in the tangent plane orthogonal to the local spin direction. This construction prevents the uncertainty model from allocating probability mass to a radial spin component that is absent from the constrained-moment supervision. Using bulk BiFeO3 and monolayer CrTe2 as benchmark systems, we show that the resulting tangent-plane epistemic uncertainty indicator Uepisf correlates strongly with prediction error and selects more informative configurations than random sampling, simultaneously improving energy, force, and projected spin-force accuracy. These results demonstrate a physically interpretable and data-efficient route for constructing uncertainty-aware magnetic machine-learning interatomic potentials.
Versions
Related
- Does System Size Matter in Uncertainty Quantification for Machine Learning Interatomic Potentials?
- Active Learning for Interatomic Potentials: A Review of Uncertainty-Driven Data Efficiency
- PEEL: Physics-Enabled Evidential Learning for Identifiable Uncertainty in CT Imaging
- Benchmarking machine-learning interatomic potentials using classical force fields
- MLIP Detective: Active Failure Mode Discovery Beyond Benchmark Scores for Machine-Learning Interatomic Potentials
BibTeX
@misc{cheng2026tangentplane,
title = {{Tangent-Plane Evidential Uncertainty in Active Learning for Magnetic Interatomic Potentials}},
author = {Yang Cheng and Hongyu Yu and Hongjun Xiang},
howpublished = {arXiv},
year = {2026},
month = jul,
doi = {10.48550/arxiv.2605.12353},
eprint = {2605.12353},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.12353},
}
RIS
TY - UNPB
TI - Tangent-Plane Evidential Uncertainty in Active Learning for Magnetic Interatomic Potentials
AU - Yang Cheng
AU - Hongyu Yu
AU - Hongjun Xiang
PY - 2026
DA - 2026/07/25
PB - arXiv
DO - 10.48550/arxiv.2605.12353
UR - https://arxiv.org/abs/2605.12353
AB - Magnetic interatomic potentials need to account for coupled lattice and spin degrees of freedom, yet constructing reliable training sets remains costly because noncollinear first-principles labels are expensive. Active learning can mitigate this cost, provided that the uncertainty estimate is physically meaningful for the magnetic-response targets that drive spin reorientation. Here we extend the e^2IP evidential framework to magnetic machine-learning interatomic potentials by formulating the projected spin-force likelihood and the corresponding epistemic uncertainty in the tangent plane orthogonal to the local spin direction. This construction prevents the uncertainty model from allocating probability mass to a radial spin component that is absent from the constrained-moment supervision. Using bulk BiFeO_3 and monolayer CrTe_2 as benchmark systems, we show that the resulting tangent-plane epistemic uncertainty indicator U_epi^sf correlates strongly with prediction error and selects more informative configurations than random sampling, simultaneously improving energy, force, and projected spin-force accuracy. These results demonstrate a physically interpretable and data-efficient route for constructing uncertainty-aware magnetic machine-learning interatomic potentials.
ER -