IBM and NASA have partnered to develop an AI foundation model for weather and climate applications, aiming to improve forecasting accuracy and provide greater accessibility and diversity of data.
Current AI models such as GraphCast and Fourcastnet can generate weather forecasts quickly but are considered AI emulators and lack the ability to encode the physics at the core of weather forecasting.
IBM and NASA's collaboration builds on their previous deployment of a foundational model for geospatial intelligence, utilizing NASA satellite data.
[
Collection
]
[
|
...
]