News: NASA has open sourced its AI system ExoMiner++ while analysing TESS data to support transparent, collaborative, and faster global exoplanet discovery.
About ExoMiner++
- ExoMiner++ is a deep learning–based artificial intelligence system used to identify exoplanets from space mission data.
- Developed by: It was developed by NASA.
- The main goals of the ExoMiner project are:
- Perform classification of transit signals in Kepler and Transiting Exoplanet Survey Satellite (TESS) data;
- Create vetted catalogs of Threshold Crossing Events (TCEs) for Kepler and TESS runs for the exoplanet community.
- Validate new exoplanets using Kepler and TESS data.
- Mechanism used: ExoMiner++ uses deep learning to study stellar light curves and detect small dips in brightness caused by planetary transits.
- Key features
- Explainable AI (XAI): The open-source system allows researchers to study, audit, and understand the model’s working.
- False Positive Mitigation: It distinguishes genuine planetary signals from eclipsing binary stars and astrophysical noise.
- Efficiency: It flagged approximately 7,000 potential exoplanet candidates from TESS data, speeding up exoplanet discovery.
- Open Science: NASA has open sourced ExoMiner++, enabling global collaboration and replication of results.
About ExoMiner
- ExoMiner is an open-source artificial intelligence software designed to analyse exoplanet data from space missions.
- Developed by: It was developed by a team from NASA’s Ames Research Center in 2021.
- The team later created ExoMiner++, trained using Kepler and TESS data.
