Machine learning helps astronomers detect new alien planets

According to the researchers, the study was published in astrophysics journal It represents the first step toward using machine learning to identify potentially previously overlooked exoplanets.

“We used traditional methods to identify the planets, but our model told us to run those simulations and show us exactly where the planets are,” the study said. Lead author Jason Terry said in a press statement: Terry is a PhD student at the University of Georgia.

When researchers applied machine learning models to a series of old astronomical observations, the models suggested the existence of planets. This was shown by a number of images highlighting specific parts of the “protoplanetary disk”, the spinning disk of dense gas surrounding the newborn star.

In that region of the disk, researchers say, there were unusual fluctuations in gas velocity. This is a characteristic sign of the planet.

According to Cassandra Hall, until now machine learning was only used to find known exoplanet formations. Researchers have now proven that the model can be used to make entirely new discoveries.

In this case, the model was able to detect signals in data that people had already analyzed. Something not detected by other researchers. According to Terry, this generally means that machine learning has the ability to quickly and accurately identify important information that people might miss.