Powerful AI finds 100+ hidden planets in NASA data including rare and extreme worlds

Astronomers at the University of Warwick confirmed more than 100 exoplanets, including 31 new ones, using a new AI system called RAVEN.
The researchers analyzed data from NASA’s TESS mission that captured observations of over 2.2 million stars in the first four years of TESS, focusing on planets orbiting very close to their stars.
The AI-based RAVEN pipeline validated 118 new planets and over 2,000 high-quality planet candidates, nearly 1,000 of them entirely new.
RAVEN helps distinguish genuine planet signals from false signals like eclipsing binary stars through a dataset of realistically simulated planets and machine learning models.
The researchers produced one of the most precise measurements of how common short-period planets are around Sun-like stars, finding 9-10% host such close-in planets.
The study also made the first direct measurement of how rare “Neptunian desert” planets are, finding they appear around only 0.08% of Sun-like stars.
The results indicate that TESS is as effective as, and sometimes better than, NASA’s Kepler mission for studying planetary populations.
RAVEN aims to improve discoverability of new planets while producing reliable datasets for broader astronomical research, and the team released interactive tools for further exploration.