Paul Allen’s AI nonprofit unveils satellite data platform

The Allen Institute for Artificial Intelligence has launched the OlmoEarth Platform, an open AI platform that aims to make satellite imagery and other data about the earth more available and useful. The platform is backed by a family of AI foundation models trained on roughly 10 terabytes of data derived from millions of observations of the planet. The OlmoEarth models can be fine-tuned for specific purposes, such as detecting changes in vegetation, with the help of a companion software tool called OlmoEarth Studio. The platform has already been used by various organizations, including the International Food Policy Research Institute and Amazon Conservation, to update maps of crops and track deforestation. The project grew out of a realization that while AI can help put earth imagery and data to use, deploying the technology can be a challenge for many organizations. To fill this gap, Ai2 created state-of-the-art models and companion tools, including OlmoEarth Studio, which simplifies the process of training the models for specific tasks. The platform allows users to analyze areas of the earth at a particular moment in time and publish or privately share maps illustrating their findings. The organization has released documentation and examples on its GitHub page, along with existing fine-tuned models that can be used or run offline on an organization's own computers. The goal of OlmoEarth is to give organizations free technology that compares favorably to existing commercial and academic AI projects, letting them efficiently analyze and visualize planetary data. Overall, the platform has the potential to accelerate the mission of various organizations working to preserve and protect the environment, and Ai2 encourages organizations that think the technology could be useful to reach out and explore how OlmoEarth can help them.