A Riemannian Framework for Learning Reduced-order Lagrangian Dynamics

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Friedl, Katharina, Jaquier, Noémie, Lundell, Jens, Asfour, Tamim, Kragic, Danica
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917939834257408
author Friedl, Katharina
Jaquier, Noémie
Lundell, Jens
Asfour, Tamim
Kragic, Danica
author_facet Friedl, Katharina
Jaquier, Noémie
Lundell, Jens
Asfour, Tamim
Kragic, Danica
contents By incorporating physical consistency as inductive bias, deep neural networks display increased generalization capabilities and data efficiency in learning nonlinear dynamic models. However, the complexity of these models generally increases with the system dimensionality, requiring larger datasets, more complex deep networks, and significant computational effort. We propose a novel geometric network architecture to learn physically-consistent reduced-order dynamic parameters that accurately describe the original high-dimensional system behavior. This is achieved by building on recent advances in model-order reduction and by adopting a Riemannian perspective to jointly learn a non-linear structure-preserving latent space and the associated low-dimensional dynamics. Our approach enables accurate long-term predictions of the high-dimensional dynamics of rigid and deformable systems with increased data efficiency by inferring interpretable and physically-plausible reduced Lagrangian models.
format Preprint
id arxiv_https___arxiv_org_abs_2410_18868
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Riemannian Framework for Learning Reduced-order Lagrangian Dynamics
Friedl, Katharina
Jaquier, Noémie
Lundell, Jens
Asfour, Tamim
Kragic, Danica
Machine Learning
By incorporating physical consistency as inductive bias, deep neural networks display increased generalization capabilities and data efficiency in learning nonlinear dynamic models. However, the complexity of these models generally increases with the system dimensionality, requiring larger datasets, more complex deep networks, and significant computational effort. We propose a novel geometric network architecture to learn physically-consistent reduced-order dynamic parameters that accurately describe the original high-dimensional system behavior. This is achieved by building on recent advances in model-order reduction and by adopting a Riemannian perspective to jointly learn a non-linear structure-preserving latent space and the associated low-dimensional dynamics. Our approach enables accurate long-term predictions of the high-dimensional dynamics of rigid and deformable systems with increased data efficiency by inferring interpretable and physically-plausible reduced Lagrangian models.
title A Riemannian Framework for Learning Reduced-order Lagrangian Dynamics
topic Machine Learning
url https://arxiv.org/abs/2410.18868