Modeling Spatially Varying Physical Dynamics for Spatiotemporal Predictive Learning

Yijun Lin, Yao Yi Chiang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Scopus citations

Abstract

Recent advances in incorporating physical knowledge into deep neural networks can estimate previously unknown governing partial differential equations (PDEs) in a data-driven way. They have shown promising results in spatiotemporal predictive learning. However, these methods typically assume universal governing PDEs across space, which is impractical for modeling complex spatiotemporal phenomena with high spatial variability (e.g., climate). Also, they cannot effectively model the evolution of potential errors in estimating the physical dynamics over time. This paper introduces a physics-guided neural network, SVPNet, which learns effective physical representations by estimating the error evolution in physics states for correction and modeling spatially varying physical dynamics to predict the next state. Experiments carried out in four scenarios, including benchmarks and real-world datasets, show that SVPNet outperforms state-of-the-art methods in spatiotemporal prediction tasks for natural processes and significantly improves prediction when training data are limited. Ablation studies also highlight that SVPNet is powerful in capturing physical dynamics in complex physical systems.

Original languageEnglish (US)
Title of host publication31st ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, ACM SIGSPATIAL GIS 2023
EditorsMaria Luisa Damiani, Matthias Renz, Ahmed Eldawy, Peer Kroger, Mario A. Nascimento
PublisherAssociation for Computing Machinery
ISBN (Electronic)9798400701689
DOIs
StatePublished - Nov 13 2023
Event31st ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, ACM SIGSPATIAL GIS 2023 - Hamburg, Germany
Duration: Nov 13 2023Nov 16 2023

Publication series

NameGIS: Proceedings of the ACM International Symposium on Advances in Geographic Information Systems

Conference

Conference31st ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, ACM SIGSPATIAL GIS 2023
Country/TerritoryGermany
CityHamburg
Period11/13/2311/16/23

Bibliographical note

Publisher Copyright:
© 2023 ACM.

Keywords

  • physics-guided machine learning
  • spatial AI
  • spatiotemporal predictive learning

Fingerprint

Dive into the research topics of 'Modeling Spatially Varying Physical Dynamics for Spatiotemporal Predictive Learning'. Together they form a unique fingerprint.

Cite this