Learning Trajectories of Hamiltonian Systems with Neural Networks

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

9 Downloads (Pure)


Modeling of conservative systems with neural networks is an area of active research. A popular approach is to use Hamiltonian neural networks (HNNs) which rely on the assumptions that a conservative system is described with Hamilton's equations of motion. Many recent works focus on improving the integration schemes used when training HNNs. In this work, we propose to enhance HNNs with an estimation of a continuous-time trajectory of the modeled system using an additional neural network, called a deep hidden physics model in the literature. We demonstrate that the proposed integration scheme works well for HNNs, especially with low sampling rates, noisy and irregular observations.
Original languageEnglish
Title of host publicationArtificial Neural Networks and Machine Learning – ICANN 2022
EditorsElias Pimenidis, Mehmet Aydin, Plamen Angelov, Chrisina Jayne, Antonios Papaleonidas
Place of PublicationBristol, UK
Number of pages12
ISBN (Electronic)978-3-031-15919-0
ISBN (Print)978-3-031-15918-3
Publication statusPublished - 7 Sept 2022
MoE publication typeA4 Conference publication
EventInternational Conference on Artificial Neural Networks - University of the West of England (UWE Bristol), Bristol, United Kingdom
Duration: 6 Sept 20229 Sept 2022

Publication series

NameLecture Notes in Computer Science
ISSN (Print)0302-9743


ConferenceInternational Conference on Artificial Neural Networks
Abbreviated titleICANN
Country/TerritoryUnited Kingdom
Internet address


  • Conservative systems
  • Deep Hidden Physics Models
  • Dynamical systems
  • Hamiltonian Neural Networks
  • Physics-Informed Neural Networks


Dive into the research topics of 'Learning Trajectories of Hamiltonian Systems with Neural Networks'. Together they form a unique fingerprint.

Cite this