Differences

This shows you the differences between two versions of the page.

--- tutorials [2021/04/26 10:37] – admin
+++ tutorials [2022/04/21 08:42] (current) – admin
@@ Line 1: / Line 1: @@
-====== Tutorials related to machine learning ======
+====== ML pipeline: Automated detection and classification of ICMEs ======
-Here, we will list either links to already existing machine learning tutorials that are very useful to start with the topic, or we provide our own tutorials mostly dealing with machine learning useful for our science cases.
+We have a [[https://github.com/epn-ml/Tutorial_IWF-ICMEs|tutorial about the ML pipeline]] developed for our [[:science_cases:iwf_science_case|IWF ICME science case]] on our [[https://github.com/epn-ml|GitHub account]].
-====== Other useful tutorials and links ======
+In the tutorial, we will introduce an ML pipeline for the automated detection of interplanetary coronal mass ejections (ICMEs) in solar wind time series data. We will guide the reader through the developed ML code with the help of a sample data set of solar wind time series data from different spacecraft (WIND, STEREO-A and STEREO-B).
-  * A very nice and comprehensice "Space Science with Python" tutorial: [[https://github.com/ThomasAlbin/SpaceScienceTutorial|https://github.com/ThomasAlbin/SpaceScienceTutorial]]
+We propose a pipeline using a UNet including residual blocks, squeeze and excitation blocks, Atrous Spatial Pyramidal Pooling (ASPP) and attention blocks, similar to the [[https://arxiv.org/pdf/1911.07067.pdf|ResUNet++]], for the automatic detection of ICMEs. The original model was used for medical image segmentation, while we are dealing with time series and therefore face a slightly different use case.
+Feel free to try out our pipeline and provide feedback about it! :)