ExoNet deep learning model identifies high-confidence TESS exoplanet candidates including habitable zone planets
Updated
Updated · Astrobiology News · Apr 22
ExoNet deep learning model identifies high-confidence TESS exoplanet candidates including habitable zone planets
7 articles · Updated · Astrobiology News · Apr 22
ExoNet analyzed 200 unconfirmed TESS planet candidates and found several high-confidence exoplanets, some located within the habitable zone, using multimodal data and advanced neural network architectures.
Trained on labeled Kepler data, ExoNet integrates phase-folded light curves and stellar parameters with multi-head attention, demonstrating strong classification performance and effective generalization to TESS observations.
This approach addresses the limitations of manual vetting for thousands of TESS candidates, highlighting the growing role of machine learning in accelerating exoplanet discovery and validation.
Trained on old Kepler data, how can we trust this AI isn't missing unique planets found only by the TESS mission?
AI models like ExoNet and RAVEN are racing to find planets. Which method will ultimately lead us to another Earth?
ExoNet just found 52 new 'habitable' worlds. How soon could the JWST tell us if any of them actually host life?
With AI pinpointing dozens of Earth-like worlds, has SETI finally been given a treasure map to find E.T.?
With AI now vetting thousands of planets, what new biases might we be introducing into our search for alien worlds?
As AI automates planet discovery, is the traditional role of the observational astronomer becoming obsolete?