DeepEM: Demonstrating a Deep Learning Approach to DEM Inversion
Creators
- 1. University of Glasgow
- 2. Lockheed Martin
- 3. University of Amsterdam
- 4. New York University
- 5. University of Oxford
- 6. Southwest Research Institute
- 7. University of Michigan
Description
DeepEM is a (supervised) deep learning approach to differential emission measure (DEM) inversion that is currently under development on GitHub.
This first release coincides with the version of DeepEM demonstrated in Chapter 4 of the Machine Learning, Statistics, and Data Mining for Heliophysics e-book (Bobra & Mason 2018). Within the chapter (and the code provided here, DeepEM.ipynb) we demonstrate how a simple implementation of supervised learning can be used to reconstruct DEM maps from SDO/AIA data. Caveats of this simple implementation and future work are also discussed.
The Machine Learning, Statistics, and Data Mining for Heliophysics e-book can be accessed at https://helioml.github.io/HelioML/, and the interactive DeepEM notebook (Chapter 4) is located at https://helioml.github.io/HelioML/04/1/notebook.
Files
PaulJWright/DeepEM-v1.0.zip
Files
(135.2 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:705316a19f013ec66fbaf596ea436398
|
135.2 MB | Preview Download |
Additional details
Identifiers
Related works
- Is supplement to
- https://github.com/PaulJWright/DeepEM/tree/v1.0 (URL)
References
- Bobra & Mason (2018). HelioML e-book. 10.5281/zenodo.2575738
- Hannah & Kontar (2012). Differential emission measures from the regularized inversion of Hinode and SDO data. 10.1051/0004-6361/201117576
- Cheung et al (2015). Thermal Diagnostics with the Atmospheric Imaging Assembly onboard the Solar Dynamics Observatory: A Validated Method for Differential Emission Measure Inversions. 10.1088/0004-637X/807/2/143