Maschinelles Lernen in der Erdsystemmodellierung

What it is about

You will learn data-driven and process-based approaches to modeling the Earth system and its components. In the end you will be able to name classical process-based models and various machine-learning methods for describing Earth system dynamics, apply interpretable AI methods, and critically evaluate approaches that combine process- and data-driven models.

What you will be able to do

• State the basics of Earth system modeling
• Name linear and nonlinear ML methods for Earth system dynamics
• Understand methods for explanation and interpretability of deep neural nets
• Combine and critically evaluate process-oriented and data-driven model approaches

What the module consists of

• SeminarStudents give talks, present review and research articles and discuss them critically
• ProjektarbeitDeepening of seminar content; if possible reproduction and continuation of scientific results

Teaching method

• Studierendenvorträgepromote independent working and critical appraisal of literature
• Diskussion wissenschaftlicher Publikationentrains critical evaluation of methods and results
• Projektarbeit mit Reproduktion von Ergebnissenenables practical deepening and application of the discussed methods
• digitale Lehrveranstaltungenables flexible, location-independent teaching and learning