Thanks to AI, scientists discover over 10,000 new planets

Thanks to a new algorithm that analyzed data from over 80 million stars, astronomers have discovered over 10,000 potential exoplanets. The program revealed hidden patterns inaccessible through conventional research, reports MIR24.

Using artificial intelligence, scientists studied the brightness fluctuations of 83,717,159 stars imaged by NASA's Transiting Exoplanet Survey (TESS) satellite. By observing subtle changes in the stars' brightness, astronomers can determine when a planet passes in front of them, briefly blocking their light. This allowed them to identify 11,554 exoplanet candidates.

Through further observations, scientists confirmed the existence of a "hot Jupiter," scientifically named TIC 183374187 b. It orbits a star 3,950 light-years from Earth.

The study covered dim stars. Due to their dim light, detecting signs of planetary transits was extremely difficult. But a new machine-learning algorithm learned to recognize subtle signs of planetary presence. The computer program also allowed the team to analyze a massive data set that humans would otherwise be unable to understand on their own.

Planets orbit their stars in periods ranging from 0.5 to 27 Earth days. They are located too close to their stars to be habitable.

Currently, the existence of more than 6,000 planets outside the Solar System has been proven. If the algorithm's findings are confirmed, that number could rise to 18,000, the scientists noted.