Foundation Model for Lunar Science

Overview

Development stage

In development or acquisitionSource: a) Pre-deployment – The use case is in a development or acquisition status.

High-impact designation

Not high-impactSource: c) Not high-impact

Impact justification

Not reported

Start date

Not reported

Withheld from public reporting

NoSource: a) No

Mission

Topic area

Science

Operational functions

No function tag matched Derived by keyword rules; see methodology

Problem the AI is intended to solve

The purpose  of Lunar FM is to overcome the limitations of traditional, task-specific Machine Learning (ML) models in analyzing the vast, diverse, and long-term datasets collected by the Lunar Reconnaissance Orbiter (LRO) mission. This includes addressing data challenges such as heterogeneity in spatial and temporal resolution, differences in data formats and calibration standards, and the significant lack of large, accurately labeled datasets (sparse ground truth) for supervised learning.

Expected benefits

The LRO Foundation Model (FM) is useful because it offers key advantages over traditional ML: Generalization: It can learn general-purpose representations that generalize better across different tasks and even data from different instruments with minimal fine-tuning. Low-Data Regimes: It excels in transfer learning, allowing knowledge from pretraining on massive unlabeled datasets to be effectively applied to downstream tasks that only have limited labeled data. Efficiency: It significantly reduces the dependence on time-consuming manual data labeling by using self-supervised pretraining on vast amounts of unlabeled LRO data. Scientific Advancement: It enables understanding of lunar features by jointly analyzing multiple data modalities, leading to new scientific insights and supporting mission-critical activities like landing site selection.

System outputs

Foundation Model: Model that can be adapted for many different applications Representations: Rich, general-purpose, and robust representations of lunar surface features and spatial relationship

Technology

AI classification

Other

System name(s)

Not reported

Custom-developed code

Not reported

Public source code

Not reported

Sourcing

How it was built

Not reported

Vendor (as reported)

Not reported

Vendors (standardized)

None on the standard list

A vendor is the supplier named by the agency. It does not identify the underlying model or AI technology.

Data and privacy

Involves PII

Not reported

Privacy Impact Assessment

Not reported

Authorization to Operate

Not reported

Demographic features

Not reported

Training and evaluation data

Not reported

Federal Data Catalog

Not reported

Governance

0 of 8 minimum-practice questions answered. A blank answer means the agency reported nothing; it does not mean the practice is absent.

Potential impacts and how they were identified

Not reported