NASA has unveiled a groundbreaking advancement in its exoplanet-hunting efforts with the introduction of ExoMiner++, an advanced artificial intelligence model. This cutting-edge technology has successfully identified over 7,000 potential exoplanet candidates in its initial application using data from the Transiting Exoplanet Survey Satellite (TESS). This remarkable achievement marks a significant leap forward in the search for planets beyond our solar system.
ExoMiner++ builds upon the foundations laid by previous efforts, notably the Kepler mission, and is trained using data from both Kepler and TESS. The initiative is part of NASA’s broader strategy to leverage open science tools, providing researchers worldwide with access to sophisticated models for planetary discovery.
An Open-source Model Trained On Two Missions
Developed by a team at NASA’s Ames Research Center in California’s Silicon Valley, ExoMiner++ represents a significant enhancement over its predecessor, ExoMiner. The original model made headlines in 2021 when it validated 370 new exoplanets using Kepler data. The upgraded version now incorporates datasets from both the Kepler and TESS missions, capitalizing on their distinct observation styles. While Kepler focused on a small region of the sky, TESS scans nearly the entire celestial dome, offering a broader scope for discovery.
The model is designed to assess transit signals, which are brief dips in a star’s brightness that may indicate a planet passing in front of it. Although not all signals point to planets—some are caused by binary stars or noise—ExoMiner++ uses deep learning to sift through massive datasets and identify the most promising candidates. The 7,000 targets flagged in the initial TESS run are now earmarked for potential follow-up by ground-based telescopes.
Open Access as a Catalyst for Discovery
A key feature of ExoMiner++ is its open-source availability. According to Kevin Murphy, NASA’s Chief Science Data Officer, “open-source software like ExoMiner accelerates scientific discovery.” The model is freely available on GitHub, allowing any qualified researcher to analyze public TESS data and search for planets.
This transparency aligns with NASA’s broader Open Science Initiative, which emphasizes the public sharing of tools, research, and results. Jon Jenkins, an exoplanet scientist at NASA Ames, explained, “Open-source science and open-source software are why the exoplanet field is advancing as quickly as it is.” The public nature of ExoMiner++ fosters collaboration and replication, both critical for scientific validation and expansion.
Preparing For A Data-rich Future
While ExoMiner++ currently requires a pre-filtered list of candidate signals to operate, developers are working on an updated version capable of detecting those signals directly from raw data. This advancement would reduce the manual workload and further streamline exoplanet discovery. Miguel Martinho, a co-investigator of ExoMiner++ and KBR employee at NASA Ames, highlighted the potential of this technology: “When you have hundreds of thousands of signals, like in this case, it’s the ideal place to deploy these deep learning technologies.”
The announcement comes as NASA continues to push the boundaries of space exploration and discovery. With the potential to revolutionize our understanding of the universe, ExoMiner++ is poised to play a pivotal role in the next phase of exoplanet research. As the model evolves, it promises to unlock new possibilities and insights into the vast cosmos that surrounds us.
Meanwhile, the scientific community eagerly anticipates the results of follow-up observations and analyses, which could confirm the existence of new worlds and further enrich our knowledge of planetary systems beyond our own.
