Application of deep learning techniques in modelling and observation of the solar photosphere

Abstract

This work is part of the applications of neural networks in the study and modeling of the phenomena present
in the solar photosphere. The proposed research is based on the network model generative adversaries using
Pytorch’s artificial intelligence modules. We aim at training a neural network capable of generating groups
of images of a high similarity with input images, These images correspond to physical magnitudes of the
solar photosphere such as density, field magnetic field, plasma velocity, temperature, among others, obtained
from the MURaM simulation code, although the neural network can be trained to generate images of any
physical magnitude. The work is focused on the generation of magnetic field images in the solar photosphere.
Results of the neural network training process are presented, as well as the comparison between the training
and generated images, and the challenges to use these tools in the study of the solar photosphere.