Prime Highlights-
- IBM and NASA released a free, open-source AI model to help scientists analyze lunar surface data more accurately.
- The model identified key lunar features up to 23% more accurately than existing widely used methods.
Key Facts-
- The AI was trained on data from nine instruments across four NASA lunar missions.
- NASA’s Artemis program plans to return astronauts to the Moon in 2028.
Background-
IBM and NASA on Thursday launched an open-source AI model that will help scientists study decades of lunar data and support future plans for astronauts to live and work on the Moon.
The tool, called the NASA-IBM Lunar Foundation Model, is freely available to researchers worldwide. Developers trained it using more than 30 layers of data gathered by nine instruments across four NASA missions, including the Lunar Reconnaissance Orbiter. It joins the Prithvi family of open AI models built by IBM and NASA, which already cover weather patterns and geospatial mapping.
The model gives researchers a faster, more precise way to study the Moon. It can help identify possible ice deposits hidden in permanently shadowed craters, map safe landing zones and study volcanic surface features, work that once relied heavily on manual review of maps and images or older machine learning tools.
In benchmark testing, the model spotted key lunar surface features up to 23% more accurately than existing widely used methods, according to NASA and IBM.
Finding lunar ice matters because it points to water and oxygen, both vital for a future Moon base and for producing fuel for missions to Mars. NASA’s Artemis program aims to return astronauts to the Moon in 2028, using the mission to test technology needed for long-term lunar living and eventual travel to Mars.
