Using AI to look at the Sun

This solar flare was captured Jan. 12 by NASA's Solar Dynamics Observatory. Stanford physicists are bringing artificial intelligence techniques to the challenge of predicting such flares. Credit: NASA/SDO and the AIA; EVE; and HMI science teams

The extent to which ICT is being used is amazing, at least to me! I just discovered that scientists are using Artificial Intelligence techniques to predict solar flares.

As you know, solar flares are intense emission of energy from the Sun in form of ionising radiation. They are the consequences of twisted magnetic fields that act as lenses focussing the energy emission from the fusion processes going on inside the Sun. Problem is, we do not understand very well wha causes these twisted magnetic fields and we don't know how to predict them. We know for sure that they occur over a period of 11 years (and we at the peak of activity just now). 
Solar flares have an impact on us here on the Earth and on astronauts traveling outside the Earth protective magnetic field: strong flares (identified as M or X)  can disrupt communications and can bathe astronauts in dangerous radiations. Hence the importance of predicting them.

Physicists at Stanford have used data collected by the Solar Dynamics Observatory, SDO, a NASA satellite collecting the equivalent of 1.5TB of data each day, to train a computer for predicting the occurrence of flairs.

They have been using Artificial Intelligence techniques to have the computer analyse the data received from the SDO. They fed the computer with about 70% of the available data and used the remaining 30% to see how good it was to predict the flares. It turns out that the computer is pretty good in predicting the flares...

Notice that it is not a matter of statistical number crunching. You need to predict more than just the occurrence of a flare. You want to know where on the surface of the Sun the flare will occur and the characteristics of the magnetic field that acting like a focussing lens will direct the radiations over a very specific cone in the space, and you need to know if the Earth or the astronauts will be in that cone....

Really amazing what ICT can do. Even if I keep saying that ICT is pervasive, well, I am still surprised to see how much pervasive it is!

Author - Roberto Saracco

© 2010-2018 EIT Digital IVZW. All rights reserved. Legal notice