NASA and IBM unveil Lunar Foundation Model that cuts polar ice prediction errors
NASA and IBM released an open-source Lunar Foundation Model trained on 17 years of orbiter data to accelerate lunar science. The model ingests nearly 2 million tile bundles across 11 data types and uses lighting geometry to improve accuracy in polar ice and crater detection. The model reduced ice prediction error by up to twenty-two percent versus top baselines.
Where this came from. One outlet reported this. We read it, then wrote the sixty words above.