Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Romagnoli, Raffaele, Kar, Soummya
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2512.19291
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917163627970560
author Romagnoli, Raffaele
Kar, Soummya
author_facet Romagnoli, Raffaele
Kar, Soummya
contents This paper investigates the stability properties of neural operators through the structured representation offered by the Hybrid B-spline Deep Neural Operator (HBDNO). While existing stability-aware architectures typically enforce restrictive constraints that limit universality, HBDNO preserves full expressive power by representing outputs via B-spline control points. We show that these control points form a natural observable for post-training stability analysis. By applying Dynamic Mode Decomposition and connecting the resulting discrete dynamics to the Koopman operator framework, we provide a principled approach to spectral characterization of learned operators. Numerical results demonstrate the ability to assess stability and reveal future directions for safety-critical applications.
format Preprint
id arxiv_https___arxiv_org_abs_2512_19291
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Stability Analysis of a B-Spline Deep Neural Operator for Nonlinear Systems
Romagnoli, Raffaele
Kar, Soummya
Systems and Control
This paper investigates the stability properties of neural operators through the structured representation offered by the Hybrid B-spline Deep Neural Operator (HBDNO). While existing stability-aware architectures typically enforce restrictive constraints that limit universality, HBDNO preserves full expressive power by representing outputs via B-spline control points. We show that these control points form a natural observable for post-training stability analysis. By applying Dynamic Mode Decomposition and connecting the resulting discrete dynamics to the Koopman operator framework, we provide a principled approach to spectral characterization of learned operators. Numerical results demonstrate the ability to assess stability and reveal future directions for safety-critical applications.
title Stability Analysis of a B-Spline Deep Neural Operator for Nonlinear Systems
topic Systems and Control
url https://arxiv.org/abs/2512.19291