EASE

Cities and companies are under growing pressure to adapt to climate change while continuing to meet ambitious sustainability goals. Across the world, they are seeking more effective and innovative ways to address rising temperatures and other climate-related challenges, while building more sustainable and resilient environments for the future. 

The WEO company supports these stakeholders by delivering AI-powered analytics based on Earth Observation (EO) data, providing valuable insights into urban heat, vegetation, and land use to support climate resilience and sustainability strategies. 

During EASE, WEO and the Luxembourg Institute of Science and Technology (LIST) will enhance and scale AI-powered analytics for EO through an HPC-powered, end-to-end AI pipeline, by using MeluXina (LuxProvide).

The project will introduce two core innovations:

1) Scalable model training using distributed computing and Super-Resolution (SR), enabling WEO to eliminate reliance on high-resolution imagery for sharpening -- removing a major barrier to geographic scalability;

2) Integration of eXplainable AI (XAI) to increase trust and transparency in decision-making, with a strong focus on Super-Resolution models behaviour and outputs. 

To this end, the project team will leverage its expertise in SR and XAI to design, develop and benchmark innovative approaches to control and prevent potential hallucinations in AI models. 

The project will enhance the spatial resolution of freely available satellite data, enabling more accurate and interpretable environmental monitoring. By improving speed, reliability, and reach, it will give city planners and infrastructure managers clearer, faster information from satellite images to design greener neighbourhoods and strengthen protection against the impacts of climate change.