Learning Hamiltonians from Trajectory Data with Kernel Methods

DOI10.48550/ARXIV.2610.17835MaRDI QIDQ7395319arXivFDO

Authors David Martín de Diego, Juan-Pablo Ortega, Miguel Vaquero, Jianyu Hu, David Martin de Diego

Publication date 13 October 2026

Published in arXiv

Tldr We use kernel methods to learn Hamiltonians from discrete trajectory observations, accounting for discretization schemes used in numerical integration; we derive reconstruction bounds and a multistride correction that improves accuracy without additional data.

Full work available at URL https://doi.org/10.48550/ARXIV.2610.17835

Mathematics Subject Classification ID

Dynamical aspects of finite-dimensional Hamiltonian and Lagrangian systems (37J99) Discretization methods and integrators (symplectic, variational, geometric, etc.) for dynamical systems (37M15) Numerical solution of inverse problems involving ordinary differential equations (65L09) Numerical methods for Hamiltonian systems including symplectic integrators (65P10) Learning and adaptive systems in artificial intelligence (68T05) Hamilton's equations (70H05)

arXiv classification

math.NAmath.SG

This page was built for publication: Learning Hamiltonians from Trajectory Data with Kernel Methods

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7395319)