Lunar analysis advances as NASA and IBM release AI model

NASA accumulated unprecedented volumes of lunar observations across multiple spaceflight programs, with records from the Lunar Reconnaissance Orbiter surpassing information gathered by every other planetary mission run by the space agency combined. Lunar research continues expanding.

Evaluating this massive volume of information requires extensive processing capacity, prompting NASA and IBM to build an artificial intelligence system that the agency officially announced on September 10 to accelerate detailed lunar evaluations. The open-source NASA-IBM Lunar Foundation Model operates without licensing fees on Hugging Face for external scientists. It is open source.

More on this story: NASA and IBM deploy open lunar AI model for space exploration

“NASA has spent decades building an extraordinary scientific record of the moon, but collecting data is only part of the job,” said NASA chief science data officer and acting chief data and AI officer Kevin Murphy during the presentation of the project.

Scientists use the system to identify impact craters, locate youthful volcanic structures, and map potential water ice deposits scattered across the lunar poles. NASA confirmed these primary applications.

Murphy stated that researchers must gain simpler pathways to analyze complex observational files. NASA and IBM previously partnered on specialized computing systems that examine Earth observation and heliophysics. Both entities plan further joint scientific deployments.

Learn more: IBM and OpenAI form an alliance to bring AI to large companies

Space agencies and commercial spaceflight firms globally incorporate machine learning software to reduce operational budgets, increase productivity, and simplify analytical workloads across planetary research. Modern planetary research adopted automated workflows rapidly.

“The NASA-IBM Lunar Foundation Model shows what’s possible when we bring AI to NASA’s petabytes of scientific data,” Murphy said, adding that “that’s a real opportunity we see with AI: turning large-scale data into new discoveries.”

Engineers trained the computational architecture primarily using historical files generated by the Lunar Reconnaissance Orbiter mission. The spacecraft spent 17 years assembling high-definition photographic coverage spanning almost the entirety of the moon.