/solar_mhd_24

Primary LanguageJupyter NotebookMIT LicenseMIT

Solar MHD 2024

Repository containing the notebooks used for the lecture on the Solar MHD 2024 meeting.

Classical machine learning

The notebook can be found here. It shows the application of principal component analysis (PCA) and k-means to real observed Hinode data. It also discusses the application of a simple neural network for regression, showing the main ingredients of how to train a neural network in PyTorch.

Complex neural networks

The notebook can be found here. It shows a much more elaborate application of neural networks to observations, focusing on using implicit neural representations or neural fields for the inference of physical properties from data.

Data

Download Hinode data used during the lecture here and here