Scientists have identified over 6,000 exoplanets—planets orbiting stars beyond our Sun. Most of these were found using data from NASA’s Kepler mission and its successor, the Transiting Exoplanet Survey Satellite (TESS). Despite the wealth of information already available, many planets are still waiting to be discovered. All data from these missions is open to the public, and researchers globally have utilized it to uncover new planets using various methods.
In 2021, a team at NASA’s Ames Research Center developed ExoMiner, an open-source software that utilized artificial intelligence to validate 370 new exoplanets based on Kepler’s data. They recently upgraded this model, now known as ExoMiner++, to include both Kepler and TESS data. This new version has already flagged around 7,000 potential exoplanets from TESS. However, these candidates still need follow-up verification with other telescopes.
ExoMiner++ is available for free on GitHub, allowing anyone interested to explore TESS data. Kevin Murphy, NASA’s chief science data officer, emphasized the importance of open-source tools in research. He noted that sharing allows for quicker discoveries and deeper data exploration.
The software examines signals from TESS observations to differentiate between actual exoplanets and other astronomical phenomena, like eclipsing binary stars. Miguel Martinho, a co-investigator of ExoMiner++, mentioned that the large volume of signals makes it a prime candidate for advanced deep learning techniques.
Kepler and TESS have different observational strategies. TESS scans almost the entire sky, focusing on nearby stars, while Kepler examined a smaller area in great detail. Despite these differences, the datasets are compatible, allowing ExoMiner++ to provide solid results. Project lead Hamed Valizadegan highlighted that even with limited resources, the outcomes can be substantial.
Future updates to ExoMiner++ aim to enhance its efficiency and help in identifying exoplanets even more effectively. Currently, the model flags potential planets but will soon be able to detect these signals directly from raw data.
Looking ahead, NASA’s upcoming Nancy Grace Roman Space Telescope is set to significantly boost the data available for exoplanet research, potentially revealing tens of thousands of new transits. Just like TESS data, this information will also be publicly available. The advancements from ExoMiner could play a crucial role in analyzing this new data.
Jon Jenkins, an exoplanet scientist at NASA Ames, noted that open science initiatives foster both quality research and software development. The quick pace of progress in exoplanet studies is largely thanks to open-source efforts.
For more on NASA’s transparent approach to scientific research and data sharing, visit science.nasa.gov/open-science.
Source link
Ames Research Center, Artificial Intelligence (AI), Exoplanets, Kepler / K2, Open Science, TESS (Transiting Exoplanet Survey Satellite)
