Artificial Intelligence Examining Kepler Data Discovers 300 Exoplanets

NASA’s first dedicated exoplanet-finding ‘hunter’, the Kepler Space Telescope, while in use, has observed hundreds of thousands of stars in search of potentially habitable worlds outside our solar system. The list compiled by the telescope during these observations continues to shed light on new discoveries, even though the telescope has been out of use for years.

A new artificial intelligence algorithm called ExoMiner, which scans this data faster and more efficiently, which experts have been examining for years, has discovered more than 300 new unknown exoplanets by examining the previously compiled data of Kepler.

None of the exoplanets found have suitable conditions for life

The telescope, which stopped working in November 2018, was looking at temporary reductions in stellar brightness that could be caused by a planet passing in front of the stellar disk. However, since there is no rule that all these reductions in brightness will be due to exoplanets, NASA’s scientists had to resort to more detailed procedures to distinguish the wrong from the real. Considering that Kepler, which has detected thousands of planet candidates so far, of which 3,000 have been confirmed – which is a very large number when we consider the total number of known exoplanets to be 4,539 – also collects a lot of data during its 10 years of service, this means that when enough data is provided, it can learn and develop its abilities. It can be said that ExoMiner, an artificial intelligence that can develop

What scientists have to do is examine the Kepler data for each candidate exoplanet; was to look at the light curve and calculate how much of the planet occupies the star, and analyze how long it takes for the so-called planet to cross the star’s disk. In some cases, the observed brightness changes were unlikely to be explained by an orbiting exoplanet. The ExoMiner algorithm, which follows exactly the same process more efficiently, also added 301 exoplanets to the Kepler planet catalog. Unfortunately, none of these new confirmed planets have conditions suitable for life on them.

“When ExoMiner says something is a planet, you can be sure it’s a planet,” said Hamed Valizadegan, ExoMiner project leader and machine learning manager at the Universities Space Research Association at NASA Ames Research Center. noted that it is more reliable. Proving their skills, scientists are now using ExoMiner to help scan data from upcoming exoplanet search missions such as NASA’s current Transit Exoplanet Survey Satellite (TESS) or the European Space Agency’s Planetary Transits and Ocillations of Stars (PLATO). wants.