AI/ML for Climate Modeling

What we do

We work on AI/ML for climate in three directions:

• Understanding and better modeling the Earth system -  can machine learning help us better represent physics and improve the models?
• Accelerating research - research is time consuming and sometimes painful. Can we make it easier so scientists can spend more time doing the science?
• Knowing when - and when not - to use it - AI/ML is not always the right tool. We work to understand the tradeoffs among accuracy, interpretability, physical consistency, and cost, so we can tell when a machine-learning approach adds value and when it doesn't.