NASA has open sourced ExoMiner++, its proven AI system for scanning TESS data from new worlds, enabling researchers around the world to accelerate the discovery of exoplanets through transparent and collaborative science.

NASA has open sourced its deep learning-based artificial intelligence system ExoMiner++, which now analyzes data from the Transiting Exoplanet Survey Satellite (TESS) to identify new exoplanets, taking a major step toward transparent and collaborative space science.

ExoMiner++ has already demonstrated its effectiveness by verifying 370 exoplanets in its initial mission, establishing its reliability in separating genuine planetary signals from false positives. In the initial analysis of the TESS data, the AI ​​flagged approximately 7,000 potential exoplanet candidates, each requiring follow-up observations for confirmation.

ExoMiner++ is designed to automate one of the most time-consuming aspects of planet hunting, greatly accelerating exoplanet discovery by quickly and accurately confirming thousands of signals. The system uses deep learning to analyze a star’s light curve and detect the subtle dips in brightness that occur when a planet passes by its host star.

Importantly, ExoMiner++ distinguishes real planets from misleading signals such as eclipsing binaries and other astrophysical noise, improving the efficiency and accuracy of NASA’s detection pipeline.

By making ExoMiner++ open source, NASA is making the model available to researchers, astronomers, and developers around the world to study, audit, and build. This open access will allow the global scientific community to reproduce results, improve detection algorithms, and adapt the techniques to other astronomical datasets, strengthening collaborative planetary science and expanding our understanding of distant planetary systems.