This post is also available in: עברית (Hebrew)
This open-source, artificial-intelligence based large language model developed by NASA and IBM is expected to speed up scientific research on climate change and have practical applications like tracking and predicting floods and wildfires.
The model, called “The Geospatial Foundation Model” is the first of its kind that NASA has collaborated to build for open release, and it is currently available on the public repository of machine learning models “Hugging Face”.
According to Cybernews, the model was built using hundreds of thousands of terabytes of data collected by NASA’s Harmonized Landsat and Sentinel-2 satellites. NASA said that this model marks an important milestone in the application of AI for Earth science, and added that the technology can be employed to track changes in land use, monitor natural disasters, and predict crop yields.
Additionally, and maybe the most crucial part, it could also act as the base to analyze datasets for advancing applications of artificial intelligence in combating climate change, NASA said.
The model is apparently able to analyze geospatial data up to four times faster than the most advanced current deep learning models and does so with half as much labeled data.
IBM said in a statement- “The need to understand quickly and clearly how Earth’s landscape is changing is one reason IBM set out six months ago in collaboration with NASA to build an artificial intelligence model that could speed up the analysis of satellite images and boost scientific discovery.”
IBM further shared that as part of its artificial intelligence and data platform Watsonx, it will release a commercial version of the model later this year.