Stability of implicit neural networks for long-term forecasting in dynamical systems - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2023

Stability of implicit neural networks for long-term forecasting in dynamical systems

Abstract

Forecasting physical signals in long time range is among the most challenging tasks in Partial Differential Equations (PDEs) research. To circumvent limitations of traditional solvers, many different Deep Learning methods have been proposed. They are all based on auto-regressive methods and exhibit stability issues. Drawing inspiration from the stability property of implicit numerical schemes, we introduce a stable auto-regressive implicit neural network. We develop a theory based on the stability definition of schemes to ensure the stability in forecasting of this network. It leads us to introduce hard constraints on its weights and propagate the dynamics in the latent space. Our experimental results validate our stability property, and show improved results at long-term forecasting for two transports PDEs.
Fichier principal
Vignette du fichier
stability_neural_network_Arxiv.pdf (334.27 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04132587 , version 1 (19-06-2023)

Identifiers

Cite

Léon Migus, Julien Salomon, Patrick Gallinari. Stability of implicit neural networks for long-term forecasting in dynamical systems. ICLR 2023 Workshop on Physics for Machine Learning, May 2023, Kigali, Rwanda. ⟨hal-04132587⟩
57 View
105 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More