LmCast :: Stay tuned in

NASA-IBM Lunar Foundation open-Source Geospatial AI Model

Recorded: Sept. 19, 2026, 5:09 a.m.

Original Summarized

USRA Contributes Planetary Science Expertise to NASA-IBM Lunar Foundation Model

Skip to main content

About USRA

Media Library

About USRA

Media Library

Contact

Suraiya Farukhi, Ph.D.Director, Communications sfarukhi@usra.edu443-812-6945

Back to Newsroom

Share this page

USRA Contributes Planetary Science Expertise to NASA-IBM Lunar Foundation Model

Share on: X

Share on: Facebook

Share on: LinkedIn


Social media

Visit our LinkedIn page
(opens in new window)

Visit our Facebook page
(opens in new window)

Follow us on X
(opens in new window)

Visit our Instagram profile
(opens in new window)

Latest news

18
Sep
2026

-

General Lester Lyles Receives 2026 Lifetime Achievement Award from the Air & Space Forces Association

16
Sep
2026

-

NASA’s Lunar Reconnaissance Orbiter Discovers a New 222-m Diameter Lunar Crater

28
Aug
2026

-

Aesteroid Named After LPI's Post Doctoral Fellow Dr. Imene Kerraouch

26
Aug
2026

-

"The Science of Artemis": An Interview with David Kring on This Week in Space

25
Aug
2026

-

New Study Dates Rare Mars Meteorite to 1.2 Billion Years Old

05
Aug
2026

-

E Street Technologies to Support $69 Million Office of Naval Research STEM Contract Through Huna Research Associates Joint Venture

18
September
2026

|
08:47 AM
America/New_York

USRA Contributes Planetary Science Expertise to NASA-IBM Lunar Foundation Model

Open-source artificial intelligence model combines diverse lunar datasets to support scientific analysis of the MoonThe NASA-IBM model reproduces patterns of lunar ice prospectivity (scaled from blue to yellow), shown at four locations (left) near the Moon’s pole. Top row: reference ice prospectivity map of Mons Mouton near the lunar south pole; middle row: predictions from the ConvNeXt model; bottom row: predictions from the NASA-IBM model. The NASA-IBM model preserves many fine-scale prospectivity patterns in the reference data. Image credit: NASA/IBM ResearchWASHINGTON, D.C., — September 18, 2026.  Universities Space Research Association (USRA) contributed planetary science expertise, lunar dataset development, and scientific evaluation to the newly released NASA-IBM Lunar Foundation Model, an open-source artificial intelligence (AI) model designed to help researchers analyze the large, diverse datasets collected by lunar missions.Developed through a collaboration led by NASA and IBM Research, the NASA-IBM Lunar Foundation Model was pretrained from scratch using SomBench, a multimodal lunar dataset containing nearly two million co-registered data bundles spanning 11 modalities and two spatial scales. The model brings together complementary information about the lunar surface, including imagery, topography, illumination geometry, thermophysical properties, mineralogy, radar, gravity, and other geologic and environmental data.USRA's contribution to the project was provided by Dr. Rachel Slank, an associate scientist with USRA's Science and Technology Institute, on assignment at NASA’s Marshall Space Flight Center. She served as a planetary science subject-matter expert on the NASA-IBM Lunar Foundation Model team. Slank worked across both the science and modeling teams, helping connect lunar science priorities and the physical characteristics of planetary datasets with decisions about model development, applications, and evaluation.The NASA-IBM Lunar Foundation 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. Together, 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.A major component of this work was the collaborative development of SomBench, the dataset used both to pretrain the NASA-IBM Lunar Foundation Model and to support standardized evaluation of lunar machine learning (ML) applications.SomBench contains two complementary components: a core multi-instrument dataset that serves as the pretraining corpus for the NASA-IBM Lunar Foundation Model and a suite of application benchmarks used to evaluate ML methods on representative lunar science problems. The benchmark suite addresses three broad science themes: impact processes, volcanic history, and polar volatiles.As part of that effort, Slank led development of the high-resolution Lunar Reconnaissance Orbiter Camera (LROC) NAC crater benchmark, manually identifying more than 49,000 lunar craters. The resulting dataset uses high-resolution lunar imagery together with co-registered digital terrain models to evaluate crater detection at meter-scale resolution. She also contributed to the broader SomBench datasets and science applications and provided extensive scientific review of both the SomBench study and the NASA-IBM Lunar Foundation Model study.“One of the biggest challenges was bringing together lunar datasets that span very different instruments and spatial resolutions (1m to 20km per pixel) while still preserving the scientific value of each dataset,” said Dr. Slank. “Because I worked across both the science and modeling teams, I could help make sure those differences were considered as the data was brought together and used to train and evaluate the foundation model. I can’t wait to see what new science researchers will be able to do with the help of the NASA-IBM Lunar Foundation Model!”The NASA-IBM Lunar Foundation Model and associated datasets are available through Hugging Face.NASA announcement:https://science.nasa.gov/science-research/artificial-intelligence-lunar-foundation-model/IBM announcement:https://newsroom.ibm.com/2026-09-10-ibm-and-nasa-release-open-source-ai-model-to-support-lunar-explorationAbout USRAFounded in 1969, under the auspices of the National Academy of Sciences at the request of the U.S. Government, the Universities Space Research Association (USRA) is a nonprofit corporation chartered to advance space-related science, technology, and engineering. 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. More information about USRA is available at  www.usra.edu.


© Universities Space Research Association
425 3rd Street SW, Suite 950, Washington DC, 20024

Footer menu

Terms of Use

Privacy Policy

For Employees

TwitterFacebookLinkedIn

The Universities Space Research Association (USRA) contributed essential planetary science expertise, developed lunar datasets, and performed scientific evaluations to the NASA-IBM Lunar Foundation Model, an open-source artificial intelligence model designed to facilitate the analysis of diverse lunar mission data. This collaborative effort, spearheaded by NASA and IBM Research, focused on creating a foundation model capable of synthesizing massive, heterogeneous datasets collected from lunar exploration for scientific insight. The NASA-IBM Lunar Foundation Model was pretrained using SomBench, a multimodal lunar dataset that comprises nearly two million co-registered data bundles spanning eleven different modalities and two spatial scales. This comprehensive data integrates various aspects of the lunar environment, including imagery, topography, illumination geometry, thermophysical properties, mineralogy, radar data, gravity measurements, and other geologic and environmental information.

USRA’s involvement was anchored by contributions from planetary science subject-matter experts, such as Dr. Rachel Slank, who served on the team to connect lunar science priorities with the physical characteristics of the planetary datasets during model development and evaluation. The model was designed with a multimodal approach, enabling it to learn intricate relationships among these disparate lunar observations rather than treating each data source in isolation. It is capable of operating across both regional-scale Wide Angle Camera observations and meter-scale Narrow Angle Camera data while effectively incorporating contextual information like terrain and illumination geometry.

The performance of the NASA-IBM Lunar Foundation Model was rigorously assessed across three critical downstream benchmarks: crater detection at both regional and meter scales, the segmentation of irregular mare patches, and the regression of lunar polar ice prospectivity. The results demonstrated that the pretrained model performed equally well or better than comparison models based on ImageNet pretraining, as well as models with identical architectures initialized without lunar pretraining. Furthermore, the study highlighted significant label efficiency in crater detection, suggesting that the representations acquired through lunar pretraining can reduce the demand for task-specific labeled data in certain applications.

A critical component of this work was the development of SomBench, which functions as both the pretraining corpus for the foundation model and a system for standardizing the evaluation of machine learning methods on representative lunar science problems. SomBench encompasses a core multi-instrument dataset intended for pretraining and an application benchmark suite addressing broad science themes such as impact processes, volcanic history, and the distribution of polar volatiles. To contribute to this effort, Dr. Slank led the development of the high-resolution Lunar Reconnaissance Orbiter Camera (LROC) NAC crater benchmark, which involved manually identifying more than forty-nine thousand lunar craters using high-resolution imagery and co-registered digital terrain models to evaluate meter-scale crater detection. This work underscored the complexity of integrating datasets that span vastly different spatial resolutions, ranging from one meter to twenty kilometers per pixel, while ensuring the preservation of the intrinsic scientific value of each source. The resulting pretrained model and associated resources are publicly available, allowing the planetary science and AI communities to use them as a reusable foundation for advancing lunar research applications.