University of Cambridge > Talks.cam > Machine Learning Reading Group @ CUED > From Stochastic Interpolants to Flow Maps: Foundations, Fast Generation, and Applications to Weather & Climate Modeling

  Browse A-Z

Log in

University Account External (via Google)

🔐 Log In

Information on

Subscribing to talksdetails Finding a talkdetails Adding a talkdetails Disseminating talksdetails Help and Documentationdetails

ℹī¸ Information

From Stochastic Interpolants to Flow Maps: Foundations, Fast Generation, and Applications to Weather & Climate Modeling

Download to your calendar using vCal

If you have a question about this talk, please contact Xianda Sun .

Stochastic interpolants unify flow-based and diffusion-based generative models under a single framework, yielding both ODE and SDE samplers from one trained model. We will cover the foundations of this framework and then introduce flow map matching – a method for learning the solution operator of the generative dynamics, enabling one-step to few-step generation with a 10-20x inference speedup. We will survey large-scale applications to probabilistic weather forecasting, including systems that outperform GenCast. We will then see Meta Flow Maps and Variational Flow Maps – two recent approaches for guidance and solving inverse problems using flow maps – and discuss open problems for weather and climate modeling.

Recommended reading:

https://arxiv.org/abs/2303.08797

Additional reading:

https://arxiv.org/abs/2406.07507 https://arxiv.org/abs/2505.18825

This talk is part of the Machine Learning Reading Group @ CUED series.

This talk is included in these lists:

Note that ex-directory lists are not shown.

Other lists

Institute of Criminology, Sidgwick Avenue, Seminar room B3, CB3 9DA LR3, Inglis Building, CUED. Clinical School William Harvey Lecture Theatre

Other talks

Secrets of the Surface: The Mathematical Vision of Maryam Mirzakhani ( EDI Film Screening) Christmas Social Evening & Annual General Meeting Title to be confirmed Thelocactus Bigger Picture Talk: Developing future marine engines running on ammonia and other green fuels Reproducibility In Astrophysical Simulations: The AGORA Initiative and a 14-Year Human Experiment (In Memory of Joel Primack)
 

Š 2006-2025 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity