Particle.news
Download on the App Store

AI Reanalysis of TESS Data Flags About 10,000 New Exoplanet Candidates

Most candidates await confirmation through follow-up observations.

Overview

  • A Princeton-led team used machine learning on TESS’s first-year sky survey to identify 10,091 new multi-transit planet candidates.
  • The search scanned light curves from 83 million faint stars with Random Forest classifiers, then human reviewers checked the signals.
  • The study lists 11,554 candidates in total, including 411 single-transit events that need more passes to pin down orbits.
  • The results appear in an arXiv preprint and have not been peer reviewed, so each object remains a candidate until follow-up confirms a true planet.
  • NASA separately released a full-sky TESS mosaic through September 2025 that charts 679 confirmed planets and 5,165 earlier candidates for context.