Improving the noise estimation of latent neural stochastic differential 
equations

Heck, L.; Gelbrecht, Maximilian; Schaub, M. T.; Boers, Niklas

doi:10.1063/5.0257224

Datensatz

DATENSATZ AKTIONENEXPORT

Zur Ablage hinzufügen

Lokale TagsFreigabegeschichteDetailsÜbersicht

Freigegeben

Zeitschriftenartikel

Improving the noise estimation of latent neural stochastic differential equations

Urheber*innen

Heck, L.
External Organizations;

/persons/resource/gelbrecht

Gelbrecht, Maximilian
Potsdam Institute for Climate Impact Research;

Schaub, M. T.
External Organizations;

/persons/resource/Niklas.Boers

Boers, Niklas
Potsdam Institute for Climate Impact Research;

Externe Ressourcen

https://doi.org/10.5281/zenodo.14534737
(Ergänzendes Material)

Volltexte (frei zugänglich)

Heck_2025_063139_1_5.0257224.pdf
(Verlagsversion), 2MB

Ergänzendes Material (frei zugänglich)

Es sind keine frei zugänglichen Ergänzenden Materialien verfügbar

Zitation

Heck, L., Gelbrecht, M., Schaub, M. T., Boers, N. (2025): Improving the noise estimation of latent neural stochastic differential equations. - Chaos, 35, 6, 063139.
https://doi.org/10.1063/5.0257224

Zitierlink: https://publications.pik-potsdam.de/pubman/item/item_32452

Zusammenfassung

Latent neural stochastic differential equations (SDEs) have recently emerged as a promising approach for learning generative models from stochastic time series data. However, they systematically underestimate the noise level inherent in such data, limiting their ability to capture stochastic dynamics accurately. We investigate this underestimation in detail and propose a straightforward solution; by including an explicit additional noise regularization in the loss function, we are able to learn a model that accurately captures the diffusion component of the data. We demonstrate our results on a conceptual model system that highlights the improved latent neural SDE’s capability to model stochastic bistable dynamics.