AI Breakthrough at University of Warwick Unveils Over 100 New Exoplanets and Expands Planetary Catalog
A cutting-edge artificial intelligence system from the University of Warwick has made significant strides in the quest to discover exoplanets, revealing more than 100 newly validated planets and identifying a wealth of additional candidates from data collected by NASA’s Transiting Exoplanet Survey Satellite (TESS). This innovative tool, named RAVEN (RAnking and Validation of ExoplaNets), was used to analyze a massive database of observations encompassing over 2.2 million stars monitored during TESS’s initial four years.
Dr. Marina Lafarga Magro, a postdoctoral researcher at the university and the lead author of the study, highlighted the magnitude of the breakthrough, stating, “Our RAVEN pipeline has validated 118 new planets and identified over 2,000 robust planet candidates, almost 1,000 of which are entirely new.” The insights gained from this analysis contribute to one of the most detailed catalogs of planets that orbit closely around their stars.
TESS detects exoplanets by observing minute decreases in star brightness that occur when a planet transits in front of a star, a method known as the transit technique. However, other astrophysical events, such as eclipsing binary stars, can create similar signals, complicating the task of distinguishing true exoplanets from false positives. RAVEN tackles this challenge by employing a comprehensive analytical framework that relies on extensive libraries of both simulated planetary transit signals and potential impostors, allowing it to discern subtle differences often overlooked by existing technologies.
Dr. Andreas Hadjigeorghiou, who spearheaded the development of RAVEN, noted that the system streamlines the entire validation process, integrating detection, machine learning, and statistical validation within a single framework. This efficiency offers a distinct advantage over other tools that may only address separate components of the workflow.
Among the notable discoveries are several ultra-short-period planets, which complete orbits in less than a day. The data indicates that approximately 9 to 10 percent of stars similar to our Sun host close-orbiting planets, aligning with previous findings from NASA’s Kepler mission but with considerably reduced uncertainties. The research also provides the first definitive assessment of the so-called ‘Neptunian desert,’ a region near stars where Neptune-sized planets are remarkably scarce. The study confirms that these planets are present around just 0.08 percent of Sun-like stars.
Dr. Kaiming Cui, another postdoctoral researcher involved in the project, expressed the significance of these findings, stating, “For the first time, we can quantify the emptiness of the ‘desert.’” Furthermore, the investigation uncovered new multi-planet systems, where multiple planets orbit the same star in close proximity. The specificity of RAVEN’s output enables astronomers to advance from examining individual exoplanets to understanding broader planetary populations.
In addition to their findings, the research team has made interactive catalogs and tools available for other scientists to explore the newly identified exoplanets, facilitating the selection of potential targets for future observations. The groundwork laid by this work is expected to be advantageous for upcoming missions, including the European Space Agency’s PLATO telescope, which aims to further study exoplanetary systems.