USRA Contributes to Open-Source AI Model for Lunar Science

Universities Space Research Association (USRA) has contributed its planetary science expertise to the development of an open-source artificial intelligence (AI) model designed to analyze lunar datasets. This model, the NASA-IBM Lunar Foundation Model, was released on September 18, 2026, and is aimed at supporting scientific analysis of the Moon.

The model was developed through a collaboration led by NASA and IBM Research. It was pretrained from scratch using SomBench, a multimodal lunar dataset containing nearly two million co-registered data bundles that span 11 modalities and two spatial scales. These modalities include imagery, topography, illumination geometry, thermophysical properties, mineralogy, radar, gravity, and other geologic and environmental data.

Dr. Rachel Slank, an associate scientist with USRA's Science and Technology Institute, played a crucial role in this project. She served as a planetary science subject-matter expert on the NASA-IBM Lunar Foundation Model team and worked across both the science and modeling teams. Her contributions were instrumental in connecting lunar science priorities with decisions about model development, applications, and evaluation.

Slank's work on the project included leading the development of the high-resolution Lunar Reconnaissance Orbiter Camera (LROC) NAC crater benchmark. She manually identified more than 49,000 lunar craters, creating a dataset that uses high-resolution lunar imagery together with co-registered digital terrain models to evaluate crater detection at meter-scale resolution. Her efforts also contributed to the broader SomBench datasets and science applications, and she provided extensive scientific review of both the SomBench study and the NASA-IBM Lunar Foundation Model study.

The model was evaluated across three downstream benchmarks: crater detection at both regional and meter scales, segmentation of irregular mare patches (IMPs), and regression of lunar polar ice prospectivity. These applications assess the model's performance across a diverse range of lunar science challenges, from identifying impact features and mapping unusual volcanic landforms to integrating environmental datasets associated with the stability and potential distribution of polar volatiles.

Across all three benchmarks, the pretrained NASA-IBM Lunar Foundation Model matched or outperformed comparison models based on ImageNet pretraining, as well as an architecturally identical model initialized without lunar pretraining. The study also demonstrated particularly strong label efficiency in crater detection, suggesting that the representations learned through lunar pretraining can reduce the amount of task-specific labeled data required for certain applications.

The multimodal design of the NASA-IBM Lunar Foundation Model allows it to learn relationships among different types of lunar observations rather than treating each dataset independently. The model was designed to operate across both regional-scale Wide Angle Camera (WAC) observations and meter-scale Narrow Angle Camera (NAC) data while incorporating information such as terrain, illumination geometry, and other lunar surface properties.

By releasing the pretrained model, fine-tuning code, and benchmark datasets openly, the NASA-IBM Lunar Foundation Model team aims to provide the planetary science and AI communities with a reusable foundation for developing new lunar research applications. The model and associated datasets are available through Hugging Face.

Founded in 1969, under the auspices of the National Academy of Sciences at the request of the U.S. Government, USRA operates scientific institutes and facilities and conducts other major research and educational programs under federal funding. It engages the university community and employs in-house scientific leadership, innovative research and development, and project management expertise.