NASA and IBM expect their open-source AI model to speed up scientific research on climate change and have practical applications like tracking and predicting floods and wildfires.
The geospatial foundation model, the first that NASA has collaborated to build for open release, is available on Hugging Face, a public repository of machine learning models.
The model was built using hundreds of thousands of terabytes of data collected by NASA’s Harmonized Landsat and Sentinel-2 satellites.
It marks a “milestone” in the application of AI for Earth science, NASA said, noting the technology can be employed to track changes in land use, monitor natural disasters, and predict crop yields.
It could also play a “pivotal” role in understanding the effects of climate change, NASA said.
According to IBM estimates, the model is able to analyze geospatial data up to four times faster than state-of-the-art deep-learning models and do so with half as much labeled data.
“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 AI model that could speed up the analysis of satellite images and boost scientific discovery,” IBM said.
IBM also plans to release a commercial version of the model later this year as part of its AI and data platform watsonx.
Your email address will not be published. Required fields are markedmarked