Machine Learning and Resampling for Stochastic Parameterization with Memory
Offered By: PCS Institute for Basic Science via YouTube
Course Description
Overview
Explore machine learning and resampling techniques for stochastic parameterization with memory in this 25-minute conference talk. Delve into the world of multiscale dynamical systems, such as the climate system, and discover how data-based methods using machine learning are revolutionizing the parameterization of unresolved processes. Learn about the advantages of stochastic parameterization over deterministic approaches in accounting for uncertainty in small-scale to large-scale process feedback. Examine recent work on constructing data-based stochastic parameterizations with memory through resampling techniques, including binning and neural networks for probabilistic classification. Gain insights into overcoming the curse of dimensionality in long memory scenarios and evaluate the performance of these approaches on various test problems.
Syllabus
Daan Crommelin: Machine Learning and Resampling for Stochastic Parameterization with Memory
Taught by
PCS Institute for Basic Science
Related Courses
Dimensionality Reduction in PythonDataCamp Sparse Nonlinear Dynamics Models with SINDy - The Library of Candidate Nonlinearities
Steve Brunton via YouTube Overcoming the Curse of Dimensionality and Mode Collapse - Ke Li
Institute for Advanced Study via YouTube Emergent Linguistic Structure in Deep Contextual Neural Word Representations - Chris Manning
Institute for Advanced Study via YouTube Multilevel Weighted Least Squares Polynomial Approximation – Sören Wolfers, KAUST
Alan Turing Institute via YouTube