Working Group, Fall 2022

High-dimensional flows and approximation

Organizers: Dejan Slepčev and Bob Pego

Participants will discuss recent works related to a variety of problems in data science and statistics that involve flows in high dimensions and their computation/approximation. Many standard methods for working with flows suffer from the `curse of dimensionality' in that the number of variables needed for good approximations explodes exponentially with the dimension. The working group will focus on a variety of new approaches that overcome the curse. Concepts that may be relevant for understanding better approaches to these problems include sliced Wasserstein distance, reproducing kernel Hilbert spaces, and the random batch method.