Particle.news
Download on the App Store
5 ARTICLES
·
6h ago

AI Reanalysis of TESS Data Flags 10,091 New Exoplanet Candidates

Machine learning on 2018 TESS observations surfaced signals around faint stars that now await verification.

The 'Milky Way' is seen in the night sky around telescopes and camps of people over rocks in the White Desert north of the Farafra Oasis southwest of Cairo May 16, 2015. The White Desert, about 500 km southwest of the Egyptian capital Cairo, features limestone and chalk forms strangely shaped by the wind and sand, a terrain that gains in intensity when illuminated by the moon. Slightly to the north lies the Black Desert, given its name by the volcanic rock dolerite, similar to basalt. Four-by-four and trekking trips for tourists include Bedouin music around campfires and nights slept under a breathtaking array of stars. Picture taken May 16, 2015. REUTERS/Amr Abdallah Dalsh
Scientists have found thousands of exoplanets (planets outside our solar system) throughout the galaxy. Most can be studied only indirectly, but scientists know they vary widely, as depicted in this artist’s concept, from small, rocky worlds and gas giants to water-rich planets and those as hot as stars.
Image

Overview

  • A research team reports 10,091 possible planets found by reprocessing NASA’s TESS data with machine learning and has posted the results as a preprint on arXiv.
  • The search widened beyond the bright stars usually studied to include about 83 million much fainter stars that TESS recorded in its first year.
  • TESS spots potential worlds by watching for tiny dips in a star’s light when a planet crosses in front, a pattern known as a transit.
  • These objects are candidates rather than confirmed planets and will need peer review and follow-up checks, such as radial-velocity measurements, to rule out false signals.
  • If many are validated, the count of known exoplanets could jump well beyond today’s 6,286 confirmed worlds, and NASA’s planned 2026 Roman Space Telescope is expected to add more finds using gravitational microlensing.