ML Innovation award at NeurIPS 2025 for work on 3D cloud reconstructions.

Shirin Ermis attended the Neural Information Processing Systems conference NeurIPS 2025 in San Diego this December. She presented a spotlight talk on 3D cloud reconstructions for tropical cyclones in the Climate Change AI workshop. The work introduces a machine learning model which can predict the 3D structure and physical properties of clouds globally from geostationary satellite imagery. This significantly expands the data available on the structure of clouds, specifically for rare extreme events such as tropical cyclones. The hope is that the work will feed into weather forecasting to improve the reliability of forecasts for tropical cyclones. 

Shirin and her team won the ML Innovation award for their work - many congratulations! You can read the published work on arXiv here and a piece on Hurricane Melissa here.

This work has been enabled by Frontier Development Lab Earth Systems Lab (https://eslab.ai), a public/private partnership between the European Space Agency (ESA), Trillium Technologies, the University of Oxford and leaders in commercial AI supported by Google Cloud, Scan Computers, Nvidia Corporation, and Pasteur Labs.