Determining the present state of magnetic fields in the solar corona is vital for space weather prediction. However, there are many technical and theoretical aspects that limit our present ability to reconstruct the fields from available observations. A numerical technique is presented to attack this inverse problem, obtaining field reconstructions given polarized-light coronagraphic images. The method is extended by the use of Physics-Informed Neural Networks and a magnetohydrostatic model of the solar plasma.
Machine Learning Methods for Solar Plasma Reconstruction
Vendredi 13 novembre 2026
de
11:00 à
12:00
Bâtiment 17, Salle de conférence 35
