Google DeepMindGoogle's AI unit has introduced an AI model. Alphaearth Foundationsit consolidates large caches of Earth observation data It produces data representations that help scientists and researchers understand and monitor our planet. The company says it's the latest AI Model A “virtual satellite”-like feature that provides insight into global change.According to Google Deepmind, the model accurately and efficiently characterizes the entire Earth's land and entire coastal waters by integrating large amounts of Earth observation data into a unified digital representation or “embedded” that computer systems can easily process. This allows the model to provide scientists with a more complete and consistent picture of the evolution of our planet.
How the Alphaearth Foundations AI Model works
According to the company, the data is captured by satellites that capture information-rich images and measurements, providing scientists and experts with near-real-time views of our planet. Although data is impactful, its complexity, multimodality, and refresh rate create new challenges in connecting different data sets and using them all effectively.The AI model visualizes the rich details of the world by assigning red, green and blue colors to three of the 64 dimensions. In Ecuador, for example, the Alphaearth Foundations can detail the agricultural plot “through a permanent cloud cover.” In Antarctica, complex surfaces are mapped clearly. This is a difficult area for irregular satellite imaging. It can even reveal changes in Canada's agricultural land use that are invisible to the naked eye.Google tests Alphaearth Foundations and says it is consistently the most accurate when compared to traditional methods and other AI mapping systems. To accelerate research and unlock use cases, Google is releasing an annual collection of embedded Alphaearth Foundations as satellite embedded datasets. Google Earth Engine. The company currently uses models to generate annual embeddings, and believes its usefulness can be further amplified by combining common inference with LLM agents. Gemini future.
