Machine learning of representative features in meteorological fields (A07)

• Subject Area: Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing ; Computer Architecture, Embedded and Massively Parallel Systems
• Term: from 2015 to 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 257899354

The focus is on advanced machine learning techniques to make further progress towards uncertainty analysis and feature detection in meteorological fields. Neural networks will be used to learn mappings from spatial coordinates to distributions of prognostic variables as well as correlation structures, and thus to convey regional forecast uncertainties. Self-supervised learning and feature identification approaches will be addressed to obtain feature descriptors that build upon the existence of recuring patterns in the data, and can be used to derive advanced strategies for feature comparison and tracking.

• DFG Programme: CRC/Transregios
• Subproject of: TRR 165: Waves to Weather
• Applicant Institution: Ludwig-Maximilians-Universität München
• Project Heads: Professor Dr. Filip Sadlo, until 6/2023; Professor Dr. Rüdiger Westermann