3d render of unreal Trappist-1 exo-planets system
(This content is subject to copyright.)
Google has open-sourced their planet-hunting AI-algorithm.
In December, NASA announced that they had found two exoplanets hiding in plain sight. The discovery was made by a neural network is trained to sift through data collected from the agency’s Kepler spacecraft.
Kepler was launched in 2009 specifically to search for exoplanets in orbit around distant stars. Astronomers detect exoplanets is based on changes in the brightness of stars. If a star dims for a short period of time, it is likely that a planet is in front of it.
In four years, Kepler observed 150,000 stars, which astronomers more data than they were able to sift through. So they only focused on the 30,000 strongest signals and be able to discover of 2,500 exoplanets. But this allowed the 120,000 signals are ignored.
Google researchers trained than their AI search, by means of the 120,000 non-analyzed signals. They fed the machine of 15,000 examples of NASA confirmed exoplanet data in order to learn how to spot the characteristics of an exoplanet.
Google now has the code on Github, along with instructions on how to use it, so that the audience can try their own heavenly discovery. However, ambitious explorers have an easier time navigating through the AI if they are familiar with programming in Python and Google’s machine-learning software, TensorFlow.
“We hope that this release will prove a useful starting point for the development of similar models for other NASA missions, such as K2 (Kepler’s second mission), and the upcoming Transiting Exoplanet Survey Satellite mission,” Christopher Shallue, the lead engineer behind Google’s exoplanet AI, wrote in a blog post.
Shallue also wrote that he hopes that this will encourage a further analysis of the remaining Kepler data.
This story was previously published in the New York Post.