Scientists use AI to discover 4 smallest exoplanets closest to host stars

(ECNS) -- An international research team developed an deep learning algorithm that combines GPU phase folding and convolutional neural networks. Using this method, the team discovered five ultra-short-period planets — each smaller than Earth with orbital periods of less than one day — in stellar photometric data released by the Kepler space telescope in 2017.

Four of these planets are the smallest exoplanets ever discovered, located closest to their host stars, and are roughly the size of Mars.

This marks the first time astronomers have used artificial intelligence to simultaneously search for candidate signals and identify genuine ones. The findings were published in the international astronomy journal Monthly Notices of the Royal Astronomical Society (MNRAS).

The research team is led by Professor Ge Jian from the Shanghai Astronomical Observatory of the Chinese Academy of Sciences.