NASA and IBM Unveil Open-Source Lunar Science AI Model

NASA and IBM have joined forces to develop an open-source AI model for lunar science, aiming to unlock the vast amounts of data collected on the Moon. The dataset used to train the model comprises over 30 spatially aligned layers from nine instruments across four missions.

The model, called the NASA-IBM Lunar Foundation Model, has been trained on a unified dataset that was specifically built for this purpose. This is because lunar observations exist at different resolutions and reference frames, making it difficult to analyze them together.

The researchers have reported significant gains over the SwinV2-B baseline, including up to 22% lower error in identifying high-potential ice areas and nearly 19% better crater work at a resolution of about 100 meters using half the training data.