NASA and IBM will collaborate to make it easier for researchers to analyse and derive insights from large NASA datasets related to Earth processes. They will create two AI foundation models to analyse petabytes of text and remote-sensing data for different tasks, questions and applications.
The first one will be trained on Earth science literature. By thematically organising journals, the model will ease the search and discovery process even for non-experts. The second model will be trained on the NASA’s Harmonized Landsat Sentinel-2 HLS dataset, which consists of a record of land-use changes captured by satellites. Its applications include tracking deforestation and invasive species, coordinating response to natural disasters and monitoring mining sites.
As part of NASA’s Open-Source Science Initiative (OSSI) commitment, the scientific community will be able to benefit from these models inclusively and transparently.