Pré-Publication, Document De Travail Année : 2025

Deep Learning for Integro-Differential Modelling

Résumé

We investigate the use of deep learning technics for manipulating systems of differential and integral equations. We train models to replace classical existing algorithms in computer algebra. A new algorithm for producing interesting datasets is presented. The trained models compete with classical algorithms and have exhibited promising new examples unhandled by computer algebra methods.

Fichier principal
Vignette du fichier
article-HAL.pdf (278.02 Ko) Télécharger le fichier

Dates et versions

hal-05230281 , version 1 (29-08-2025)
hal-05230281 , version 2 (16-03-2026)
hal-05230281 , version 3 (30-03-2026)

Licence

Identifiants

  • HAL Id : hal-05230281 , version 1

Citer

Louis Roussel, François Lemaire. Deep Learning for Integro-Differential Modelling. 2025. ⟨hal-05230281v1⟩

Collections

1689 Consultations
408 Téléchargements

Partager

  • More