Dear all,
We have recently contributed a chapter based on
DeepEM: a (supervised) deep learning approach to DEM inversion. DeepEM is currently under development on Github (https://github.com/PaulJWright/DeepEM), and DeepEM
has been designed to both accelerate the computation of DEM solutions, and reduce erroneous solutions provided by
pre-existing inversion methods. The underlying idea is that a neural network should be able to learn a generalized mapping in order to transform an SDO/AIA
input vector (6x1x1) to the DEM output vector (Nx1x1, where N is the number of temperature bins of the DEM solution) also retain the fidelity to the original inversion method.
In our DeepEM chapter we provide a simple model in order to demonstrate how easy it is to apply deep learning techniques to your own research. This very simple implementation is trained on one set
of 512x512 AIA images, and corresponding Basis Pursuit DEM solutions. This implementation returns >10 Million DEM/s with similar fidelity to Basis Pursuit. A gif of how this implementation
compares to Basis Pursuit is shown here: https://twitter.com/PaulJWright/status/1068115587614863360, and
I
hope that this chapter, and e-book in general, will encourage those who are interested to applying deep learning to their research.
Cheers,
Paul
---
Paul J. Wright
Research Assistant & Ph.D. Student
Astronomy & Astrophysics Group
SUPA School of Physics & Astronomy
Kelvin Building
University of Glasgow
Glasgow, UK
G12 8QQ