NASA's open-source science initiative: A decade-long commitment to advancing knowledge

IBM and Hugging Face have recently announced a groundbreaking partnership to democratise access to AI technology in the field of climate science. The result of this collaboration is the watsonx.ai Geospatial foundation model, which stands as the largest of its kind on Hugging Face and the first-ever open-source AI foundation model developed in conjunction with NASA.

In the realm of climate science, the challenges presented by rapidly changing environmental conditions demand the latest and most comprehensive data. Despite the wealth of information available, scientists and researchers face significant obstacles in effectively analysing vast datasets.

In response, IBM initiated a Space Act Agreement with NASA to create an AI foundation model for geospatial data.

Sriram Raghavan, VP at IBM Research AI, commented:

“The essential role of open-source technologies to accelerate critical areas of discovery such as climate change has never been clearer. By combining IBM’s foundation model efforts aimed at creating flexible, reusable AI systems with NASA’s repository of Earth-satellite data, and making it available on the leading open-source AI platform, Hugging Face, we can leverage the power of collaboration to implement faster and more impactful solutions that will improve our planet.”

Sriram Raghavan, VP at IBM Research AI

Power of Geospatial Model

The Geospatial model is a product of the joint efforts of IBM and NASA, meticulously trained on Harmonised Landsat Sentinel-2 satellite data (HLS) for one year across the United States. The initial results are highly promising, demonstrating a 15 percent improvement over state-of-the-art techniques, all while utilising only half the labelled data.

The model's capabilities extend beyond its current state, as it can be further fine-tuned for diverse tasks, including deforestation tracking, crop yield prediction, and greenhouse gas detection.

NASA's decade-long open-source science initiative

IBM's partnership with NASA in building the AI model aligns with NASA's long-standing commitment to open-source science through its decade-long Open-Source Science Initiative. As part of the initiative, NASA has designated 2023 as the Year of Open Science, highlighting the importance of sharing data, information, and knowledge openly within the scientific community.

IBM’s watsonx․ai geospatial foundation model is now openly available on @huggingface. Built with @NASAEarth, it’s the first-ever open-source #AI foundation model built with NASA & largest geospatial model on HuggingFace. Explore more: https://t.co/bxRwoFAGDV pic.twitter.com/TgEEzOOE16

— IBM News (@IBMNews) August 3, 2023

The geospatial model is a testament to IBM's foundation model technology, embodying its vision of creating and training AI models with transferable capabilities across various tasks. Moreover, in June, IBM introduced ‘watsonx,’ an AI and data platform that aims to scale and accelerate the impact of advanced AI by leveraging trusted data.

Kevin Murphy, Chief Science Data Officer at NASA, said:

“We believe that foundation models have the potential to change the way observational data is analysed and help us to better understand our planet. By open-sourcing such models and making them available to the world, we hope to multiply their impact.”

Kevin Murphy, Chief Science Data Officer at NASA

Commercialisation & future prospects

A commercial version of the geospatial model, integrated into IBM watsonx, will be made available through the IBM Environmental Intelligence Suite (EIS) later this year. By harnessing the power of open-source technologies, this groundbreaking collaboration seeks to effectively address climate challenges and contribute to a more sustainable future for our planet.

Through the democratisation of AI technology and information sharing, IBM and Hugging Face are paving the way for accelerated progress in climate and Earth science, ushering in a new era of accessibility and inclusivity within the scientific community.