Regensburg 2025 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
MA: Fachverband Magnetismus
MA 35: PhD Focus Session: Using Artificial Intelligence Tools in Magnetism
MA 35.5: Talk
Thursday, March 20, 2025, 11:45–13:00, H20
Hands-on workshop for using AI in magnetism research — •Kilian Leutner1, Thomas B. Winkler2, Jan Maskill3, Kübra Kalkan4, and Robin Msiska4 — 1Institute of Physics, Johannes Gutenberg-University Mainz — 2Institute for Molecules and Materials, Radboud University — 3Department of Physics, University of Kaiserslautern-Landau — 4Faculty of Physics, University of Duisburg-Essen
In this workshop, participants will receive a practical introduction to applying artificial intelligence (AI) in magnetism research. As a hands-on session, everyone can actively engage. We will demonstrate how to automatically analyze magneto-optical Kerr microscopy images of magnetic skyrmions using AI. The critical and significant step in the analysis involves segmenting the Kerr images using the Skyrmion U-Net, a convolutional neural network [1]. The participants will use their own devices to explore a repository, learning and executing the steps necessary to train and apply a skyrmion U-Net model. The required experimental Kerr data, pre-trained Skyrmion U-Net models, and code are available in a repository at https://github.com/kfjml/AI-Magnetism-Session-Regensburg-2025 or can be downloaded as a ZIP file from: https://download.klaeui-lab.de/AI-Magnetism-Session-Regensburg-2025. Please follow the instructions in the Readme file prior to the session, to make most use of it. This presented approach, is adaptable to other magnetic textures or imaging methods and can also be applied beyond magnetism.
[1] I. Labrie-Boulay et al., Phys. Rev. Applied 21, 014014 (2024)
Keywords: Artificial intelligence; AI in research; Workshop; Kerr microscopy; Magnetism