First Time AI Successfully Reveals Photos of the Sun’s Poles, Never Seen Before

ANTARIKSA — Artificial intelligence (AI) has succeeded in revealing the first image of the sun’s poles. The solar poles are parts of the sun that have never been seen before.

So far, observations of the sun have usually been limited to viewing from the side only. Even satellite missions sent to study the sun are largely limited to studying it from an equatorial perspective.

Now, researchers are inputting two-dimensional images of the sun into artificial intelligence (AI). The image is then translated into a 3D reconstruction showing changes in the sun’s surface over time.

Once the AI ​​learned how to accurately reconstruct the past appearance of the sun for the area observed by the satellite, the team also extrapolated its polar appearance.

The results provide a more complete picture of the sun and will help scientists understand how its radiation is likely to impact power grids, satellites and radio communications on Earth.

The study was conducted by solar physicist Benoit Tremblay of the National Center for Atmospheric Research (NCAR) in Colorado and his colleagues.

“The best way to see the sun’s poles is to send more satellites, but that is very expensive,” Tremblay said in a statement, reported by Newsweek.

“By leveraging the information we have, we can use AI as a virtual observatory and give us a pretty good idea of ​​what the sun’s poles look like at a fraction of the cost.”

To develop the solar model, the researchers used artificial neural networks known as “Neural Radiance Fields” (NeRFs). NeRFs are capable of taking two-dimensional images and turning them into complex three-dimensional views.

This modeling is not easy. The first challenge arises because NeRF has never been used to image extreme ultraviolet plasmas.

Therefore, Tremblay and his colleagues had to adapt the neural network to work with solar conditions. Scientists created a new system they dubbed Sun Neural Radiance Fields, or “SuNeRFs.”

SuNeRF was trained on a time series of images taken by three satellites that have observed the sun in extreme ultraviolet light wavelengths from different angles.

The team warned that their AI-generated model of the sun was just an “intelligent guess.” However, it remains a useful estimate that can help study the sun and inform future missions sent to study the sun up close.

With the initial study complete, the team is now looking for ways to use NCAR’s Derecho supercomputer to increase the resolution of their model.