NASA and Google discovered an eight-planet circling around Kepler-90 which is Sun-like star and 2,545 light-years from the Earth. The planet was discovered with the aid of the data collected from NASA's Kepler Space Telescope.
Recently discovered Kepler-90i is a hot, rocky planet that orbits its star once in 14,4 days and it is found with the power of Machine Learning from Google. Google has been working continuously to improve machine learning algorithms and with neural networks, it has been the fastest advancement in the Machine Learning.
Researchers Christopher Shallue and Andrew Vanderbrug trained the computer to learn how to recognize the exoplanets in light readings recorded by Kepler. After the capture of the minuscule brightness when it passed in front of a star and with the help of the neural network it found weak transit signals from a previously missed eighth planet orbiting Kepler-90 in the constellation Draco. The Kepler-90 is like a mini version of our solar system, comprising small planets inside and big planets outside. Software Engineer from Google's research team Google AI came up with an idea to implement the neural network to Kepler data and became interested in the discovery of the exoplanets. Weakest signals are often missed using the methods and the scientists from NASA thought it could be more interesting in the discovery of exoplanets. Not only Kepler-90i was discovered using the neural network, Kepler-80 which was the sixth planet and Earth-sized Kepler-80g and its neighboring planets were locked by their mutual gravity in a rhythmic orbital dance.
The research paper is published of their findings and has been accepted for the publication in the Australian Journal. Kepler has produced a lot of data set for the discovery of exoplanet and applying machine learning algorithms there can be the significant advancement in the understanding of the planetary systems around the other stars.