Sequential-in-Time Training of Nonlinear Parametrizations for Solving Time-Dependent Partial Differential Equations
SIAM Review · 30.09.2026 · 68 (3) · 579-608
DOI: 10.1137/24m1651241 ↗Kurz gefasst
Fachartikel in der Zeitschrift SIAM Review, online erschienen am 06.08.2026. Verfasst von Huan Zhang und 3 weiteren Autor:innen. Die Arbeit nennt 95 Literaturangaben. Gefördert von: Army Research Office, Office of Naval Research, National Science Foundation.
Abstract (Originalsprache)
Sequential-in-time methods solve a sequence of training problems to fit nonlinear parametrizations such as neural networks to approximate solution trajectories of partial differential equations over time. This work shows that sequential-in-time training methods can be understood broadly as either optimize-then-discretize (OtD) or discretize-then-optimize (DtO) schemes, which are well-known concepts in numerical analysis. The unifying perspective leads to novel stability and a posteriori error analysis results that provide insights into theoretical and numerical aspects that are inherent to either OtD or DtO schemes such as the tangent space collapse phenomenon, which is a form of overfitting. Additionally, the unified perspective facilitates the establishment of connections between variants of sequential-in-time training methods, which is demonstrated by identifying natural gradient descent methods on energy functionals as OtD schemes applied to the corresponding gradient flows.
Förderung
- Army Research Office
- Office of Naval Research
- National Science Foundation
Zitieren
APA
Zhang, H., Chen, Y., Vanden-Eijnden, E., & Peherstorfer, B. (2026). Sequential-in-Time Training of Nonlinear Parametrizations for Solving Time-Dependent Partial Differential Equations. SIAM Review, 68(3), 579-608. https://doi.org/10.1137/24m1651241
Vancouver
Zhang H, Chen Y, Vanden-Eijnden E, Peherstorfer B. Sequential-in-Time Training of Nonlinear Parametrizations for Solving Time-Dependent Partial Differential Equations. SIAM Review. 2026;68(3):579-608. doi:10.1137/24m1651241
BibTeX
@article{Zhang2026,
author = {Zhang, Huan and Chen, Yifan and Vanden-Eijnden, Eric and Peherstorfer, Benjamin},
title = {Sequential-in-Time Training of Nonlinear Parametrizations for Solving Time-Dependent Partial Differential Equations},
journal = {SIAM Review},
year = {2026},
volume = {68},
number = {3},
pages = {579--608},
doi = {10.1137/24m1651241}
}
EUVIMED