Use Cases

Embed2Scale aims to advance the state-of-the-art in neural data compression for Earth observation. For this purpose, consortium partners assembled an Earth observation benchmark dataset to develop, test, and rank neural compression methodologies. 
The project focuses on four real‑world applications that show how AI‑compression, machine‑learning, and high‑performance computing can make Earth observation data more accessible and scalable.

The four use cases explore: 

• Maritime awareness (led by SATCEN)
• Crop Stress and Yield Early Detection (led by Jülich Supercomputing Centre)
• Global aboveground biomass estimation (led by UZH)
• Climate and air pollution prediction from spatio-temporal observational constraints (led by UOXF)”