HQ Team
August 5, 2023: IBM’s geospatial foundation model, Watson.ai, is now available in the open-source artificial intelligence platform, Hugging Face — the first such built in collaboration with NASA.
The move will enable scientists to analyze satellite images and predict trends in climate change more accurately, according to an IBM statement. It will make about 250,000 terabytes of NASA mission data accessible to more people.
The platform can analyze geospatial data up to four times faster than state-of-the-art deep-learning models, with half as much labeled data, IBM estimated.
As part of a Space Act Agreement with NASA, IBM set out earlier this year to build an AI foundation model for geospatial data.
Democratize access
Foundation models are highly versatile, and by open-sourcing, Hugging Face, NASA, and IBM hope that researchers worldwide would be “motivated to improve on it and build other geospatial models and applications.”
By making the model available via Hugging Face, a repository for all transformer models, efforts could be advanced to democratize access and application of AI to generate new innovations in climate and Earth science, according to the statement.
Access to the latest data remains a significant challenge in climate science where environmental conditions change almost daily. And, despite growing amounts of data scientists and researchers still face obstacles in analyzing these large datasets.
Flood, burn scar mapping
The model – trained jointly by IBM and NASA on Harmonized Landsat Sentinel-2 satellite data (HLS) over one year across the US — has been fine-tuned on labeled data for flood and burn scar mapping.
In burn scar mapping scientists map fire history by detecting ‘fire scars’, the visibly blackened land surface left after a fire has burned through vegetation.
It has demonstrated to date a 15% improvement over state-of-the-art techniques using half as much labeled data.
“AI remains a science-driven field, and science can only progress through information sharing and collaboration,” said Jeff Boudier, head of product and growth at Hugging Face.
“This is why open-source AI and the open release of models and datasets are so fundamental to the continued progress of AI, and making sure the technology will benefit as many people as possible.”
The base model, when fine-tuned, can be redeployed for tasks like tracking deforestation, predicting crop yields, or detecting and monitoring greenhouse gasses.
‘Multiply impact’
IBM and NASA researchers also work with Clark University to adapt the model for applications such as time-series segmentation and similarity research.
A commercial version of the geospatial model, which is part of IBM watsonx, will be available through the IBM Environmental Intelligence Suite later this year, according to the statement.
“We believe that foundation models have the potential to change the way observational data is analyzed and help us to better understand our planet,” said Kevin Murphy, Chief Science Data Officer, NASA.
“And by open-sourcing such models and making them available to the world, we hope to multiply their impact.
