Seminar AI for Science

What it is about

In this seminar you explore machine learning in natural and engineering sciences, e.g. CNNs in fluid mechanics, GNNs for molecular properties and Fourier Neural Operators in weather modeling. By the end you understand the methodology of current state-of-the-art approaches, know their strengths and limits, and can scientifically prepare a research topic, present it, and discuss it with a multidisciplinary audience.

What you will be able to do

• Understand methodology of current state-of-the-art methods in scientific ML
• Assess strengths and limits of these methods
• research and develop scientific questions
• Present results effectively and defend them in discussions
• Analyze social, economic and ecological impacts of the methods

What the module consists of

• SeminarThe central teaching format: you work independently on an assigned research topic and summarize the results in writing; you present your report and discuss it with the participants.

Teaching method

• Independent research and written elaborationto deeply research a topic and present it in writing
• Presentation with moderated discussionto convey the results, receive feedback and foster professional exchange