Inference of Hamiltonian and Lie-Poisson structures from vector field samples

DOI10.48550/arXiv.2610.04350MaRDI QIDQ7367910zbMATH OpenarXivFDO

Authors Georg A. Gottwald, Jason E. Frank

Publication date 3 October 2026

Published in arXiv

Tldr We present a method for learning Hamiltonian and Lie-Poisson dynamical systems from vector-field samples by jointly estimating the Hamiltonian and underlying Poisson tensor. We treat three cases with increasing generality, recovering the Hamiltonian and structure matrix via eigenvalue problems and alternating between generalized eigenvalue and nonlinear least-squares optimization.

Full work available at URL https://doi.org/10.48550/arXiv.2610.04350

Mathematics Subject Classification ID

General theory of finite-dimensional Hamiltonian and Lagrangian systems, Hamiltonian and Lagrangian structures, symmetries, invariants (37J06) Discretization methods and integrators (symplectic, variational, geometric, etc.) for dynamical systems (37M15) Poisson manifolds; Poisson groupoids and algebroids (53D17) Artificial neural networks and deep learning (68T07) System identification (93B30)

arXiv classification

math.DSmath.NAmath.SG

Cites work

• A framework for machine learning of model error in dynamical systems
• A Review of Data‐Driven Discovery for Dynamic Systems
• A structure-preserving kernel method for learning Hamiltonian systems
• Combining machine learning and data assimilation to forecast dynamical systems from noisy partial observations
• Discovering governing equations from data by sparse identification of nonlinear dynamical systems
• GFINNs: GENERIC Formalism Informed Neural Networks for Deterministic and Stochastic Dynamical Systems
• Lie-Poisson neural networks (LPNets): data-based computing of Hamiltonian systems with symmetries
• Simulating Hamiltonian Dynamics
• SINDy-PI: a robust algorithm for parallel implicit sparse identification of nonlinear dynamics
• Supervised learning from noisy observations: combining machine-learning techniques with data assimilation
• Symplectic learning for Hamiltonian neural networks
• The Random Feature Model for Input-Output Maps between Banach Spaces
• Universal approximation bounds for superpositions of a sigmoidal function

This page was built for publication: Inference of Hamiltonian and Lie-Poisson structures from vector field samples

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7367910)