dc.contributor.advisor |
Butz, Martin |
|
dc.contributor.author |
Karlbauer, Matthias |
|
dc.date.accessioned |
2024-08-15T07:29:17Z |
|
dc.date.available |
2024-08-15T07:29:17Z |
|
dc.date.issued |
2024 |
|
dc.identifier.uri |
http://hdl.handle.net/10900/156815 |
|
dc.identifier.uri |
http://nbn-resolving.de/urn:nbn:de:bsz:21-dspace-1568159 |
de_DE |
dc.identifier.uri |
http://dx.doi.org/10.15496/publikation-98147 |
|
dc.description.abstract |
This thesis explores the potential of machine learning methods for improving weather forecasts. Since weather is considered a spatiotemporal process that evolves over space through time, the thesis first investigates the design choices required for machine learning models to simulate synthetic spatiotemporal processes, such as the two-dimensional wave equation. It then develops a method for analyzing machine learning models that enables the extraction of unknown process-relevant context that parameterizes an observed simulated spatiotemporal process of interest. Relating these extracted factors to physical properties leads the thesis to physics-aware machine learning, where it explores how to fuse process knowledge from physics with the learning ability of artificial neural networks. Given the insights from those investigations, a competitive deep learning weather prediction model is designed to understand which design choices support data-driven algorithms to learn a meaningful function that predicts realistic and stable states of the atmosphere over hundreds of hours, days, and weeks into the future. |
en |
dc.language.iso |
en |
de_DE |
dc.publisher |
Tübingen Library Publishing |
de_DE |
dc.rights.uri |
https://creativecommons.org/licenses/by/4.0/deed.de |
|
dc.subject.classification |
Deep learning , Wetter , Wettervorhersage , Maschinelles Lernen , Neuronales Netz |
de_DE |
dc.subject.ddc |
004 |
de_DE |
dc.subject.other |
spatiotemporal dynamics |
en |
dc.subject.other |
knowledge extraction |
en |
dc.subject.other |
physics aware neural networks |
en |
dc.subject.other |
deep learning |
en |
dc.subject.other |
weather prediction |
en |
dc.subject.other |
deep learning weather prediction |
en |
dc.title |
Artificial Neural Networks for Knowledge Extraction in Spatiotemporal Dynamics and Weather Forecasting |
en |
dc.type |
PhDThesis |
de_DE |
dcterms.dateAccepted |
2024-03-22 |
|
utue.publikation.fachbereich |
Informatik |
de_DE |
utue.publikation.fakultaet |
7 Mathematisch-Naturwissenschaftliche Fakultät |
de_DE |
utue.publikation.noppn |
yes |
de_DE |