Solar flares can be dangerous. The release of energy from a flare is like detonating a mass of atomic bombs, which can damage power grids on Earth 93 million miles away and disturb satellite communication. Scientists do not know for certain what triggers solar flares, according to a press release from the Stanford News Service, but physicists Monica Bobra and Sebastien Couvidat have employed artificial intelligence to help predict the energy surges.
A solar flare happens when magnetic energy on the sun's corona gets tangled and snaps. The sun is now at the peak of an 11-year cycle, during which the flares occur more frequently.
Bobra and Couvidat of Stanford University used A.I. performance to automate the analysis of solar observations from the Solar Dynamics Observatory (SDO) in order to ascertain which features aid in solar flare prediction.
Instruments that measure vector magnetic field data usually only capture what is in the line of sight, therefore, only part of the sun can be studied part of the time. Helioseismic Magnetic Imager (HMI) is now a part of SDO's arsenal for solar observation and it can collect data almost continuously, according to the press release.
"Machine learning is a sophisticated way to analyze a ton of data and classify it into different groups," Bobra said.
"It's exciting because we not only have a ton of data, but the images are just so beautiful," she later said, according to the press release. "And it's truly universal. Creatures from a different galaxy could be learning these same principles."
