A machine learning analysis of NASA’s TESS data has identified thousands of new exoplanet candidates.
A study uploaded to arXiv on April 20 reports the identification of 11,554 exoplanet candidates through a machine learning analysis of data from NASA's Transiting Exoplanet Survey Satellite (TESS). Of these, 10,052 are newly identified candidates.
Key Findings
The analysis detected 11,554 candidates, of which 10,052 had not been previously reported. Researchers analyzed light curves of 83,717,159 stars from TESS's first wide-field image.
Approximately 87% of the candidates show two or more transits, allowing calculation of orbital periods ranging from 0.5 to 27 days.
One candidate, TIC 183374187 b, a hot Jupiter orbiting a star about 3,950 light-years away, was confirmed using the Magellan telescopes in Chile.
The brief orbital periods suggest these planets are likely too close to their stars to support life as known. The newly identified candidates orbit faint stars, up to 16 magnitudes dimmer than typical transit survey thresholds.
Context
As of September 2025, the number of confirmed exoplanets exceeded 6,000. If all candidates are confirmed, the total could reach nearly 18,000. Confirmation requires independent surveys and further study, which may take months or years.
Source
The study has not yet been peer-reviewed. It was uploaded to the preprint server arXiv on April 20, 2025.