Over 100 hidden planets discovered in NASA data: Artificial intelligence at the service of astronomy

An international team of astronomers validated over 100 hidden exoplanets using the RAVEN AI tool developed at the University of Warwick in the United Kingdom.
The system analyzed data from NASA’s TESS, reviewing observations of 2.2 million stars over four years.
Among their findings, they identified 31 previously undiscovered planets and over 2,000 high-quality candidates.
The team identified ultra-short period exoplanets with orbits lasting less than 24 hours and detected planets in the “Neptunian desert.”
Validation included precise characterization and multiplanetary systems with previously undetected bodies, managed through a well-characterized sample of planets near Earth.
RAVEN distinguishes exoplanets from false signals by employing a large catalog of simulated planets and astrophysical phenomena to identify complex patterns.
The study found that between 9 and 10% of solar-type stars host short-orbit exoplanets and only 0.08% of solar stars contain planets in the “Neptunian desert.”
The team has made public catalogs and interactive tools available for other researchers, enabling future observations via ground-based telescopes and ESA’s PLATO space mission.

You may have missed