EPS Colloquium – Peter Battaglia, Google DeepMind

AI for Weather

This talk will describe a program of research on developing AI models for weather forecasting. We’ll start with early work on machine learning for simulation, and explore how these early efforts paved the way for the rapidly advancing field of models which now far outperform traditional weather forecasts. We’ll also talk about how to make AI models for science most useful to people, and the future of the field more broadly.

To be added to the EPS colloquium mailing list, please contact Caroline Carr at carolinecarr@fas.harvard.edu

Peter Battaglia is a Senior Director of Research at Google DeepMind, where he leads initiatives at the intersection of AI and environmental sustainability. His work is best known for advancing the use of graph neural networks and structured models to enable machines to reason about physics and complex systems. Battaglia’s team developed WeatherNext, Google’s family of state-of-the-art AI models for weather forecasting. His broader portfolio also includes energy, agriculture, and biodiversity AI research programs. His aim is to make transformative AI innovations, and apply them to solving the most critical real-world challenges.