IBM, NASA launch AI model to help map ice, craters on Moon

NASA and IBM have unveiled a groundbreaking open-source AI model designed to aid future moon explorations as part of NASA's Artemis program. The model, named the NASA-IBM Lunar Foundation Model, is now accessible on the collaborative Hugging Face platform. Trained on decades of lunar observation data, the model offers researchers a powerful tool to extract valuable insights into the moon's geography, including craters, volcanic formations, and potential ice deposits.

By automating the process of data analysis, the model saves valuable time and outperforms previous methods by up to 23%. The project builds upon a decade-long collaboration between NASA and IBM, aiming to foster a shared foundation of knowledge for the global scientific community. The model and its associated open-source lunar dataset, comprised of tens of thousands of maps and images from nine instruments across four moon missions, are now available for researchers to adapt and apply to new questions about the moon.