Weighted Birkhoff averages accelerate data-driven methods
Proceedings of the Royal Society A Mathematical Physical and Engineering Science · 01.03.2026 · 482 (2333)
DOI: 10.1098/rspa.2025.0979 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the Royal Society A Mathematical Physical and Engineering Science, online erschienen am 11.03.2026. Verfasst von Maria Bou-Sakr-El-Tayar und 2 weiteren Autor:innen. Die Arbeit nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Natural Sciences and Engineering Research Council of Canada, Isaac Newton Trust, Fonds de Recherche du Québec Nature et Technologies.
Abstract (Originalsprache)
Many data-driven algorithms in dynamical systems rely on ergodic averages that converge painfully slowly. One simple idea changes this: taper the ends. Weighted Birkhoff averages can converge much faster (sometimes superpolynomially, even exponentially) and can be incorporated seamlessly into existing methods. We demonstrate this with five weighted algorithms: weighted dynamic mode decomposition (wtDMD), weighted extended DMD (wtEDMD), weighted sparse identification of nonlinear dynamics (wtSINDy), weighted spectral measure estimation and weighted diffusion forecasting. Across examples ranging from fluid flows to El Niño data, the message is clear: weighting costs nothing, is easy to implement and often delivers markedly better results from the same data.
Förderung
- Natural Sciences and Engineering Research Council of Canada
- Isaac Newton Trust
- Fonds de Recherche du Québec Nature et Technologies
Zitieren
APA
Bou-Sakr-El-Tayar, M., Bramburger, J. J., & Colbrook, M. (2026). Weighted Birkhoff averages accelerate data-driven methods. Proceedings of the Royal Society A Mathematical Physical and Engineering Science, 482(2333). https://doi.org/10.1098/rspa.2025.0979
Vancouver
Bou-Sakr-El-Tayar M, Bramburger JJ, Colbrook M. Weighted Birkhoff averages accelerate data-driven methods. Proceedings of the Royal Society A Mathematical Physical and Engineering Science. 2026;482(2333). doi:10.1098/rspa.2025.0979
BibTeX
@article{BouSakrElTayar2026,
author = {Bou-Sakr-El-Tayar, Maria and Bramburger, Jason J. and Colbrook, Matthew},
title = {Weighted Birkhoff averages accelerate data-driven methods},
journal = {Proceedings of the Royal Society A Mathematical Physical and Engineering Science},
year = {2026},
volume = {482},
number = {2333},
doi = {10.1098/rspa.2025.0979}
}
EUVIMED