Wasserstein Gradient Flows in Math and Machine Learning (25w5430)


Soumik Pal (University of Washington, Seattle)

Young-Heon Kim (University of British Columbia)

Anna Korba (ensae/crest)

Brendan Pass (University of Alberta)


The Banff International Research Station will host the “Wasserstein Gradient Flows in Math and Machine Learning” workshop in Banff from June 29 - July 4, 2025.

Much of modern machine learning is based on optimization where one tries to find parameters of a model that minimize loss. These optimization problems are then solved numerically by running a dynamic algorithm. One of the major theoretical difficulties is the scale of these problems. Modern data are huge in both their dimension and their count. A proper understanding requires a mathematical theory that is well suited to understand the effect of dimensions on these problems. Wasserstein gradient flow is one such mathematical theory. Originally developed for studying partial differential equations, in recent years it has appeared to have the right tools to study a number of popular machine learning and data science algorithms. The aim of this workshop is to bring together prominent researchers from both mathematics and machine learning, along with a number of young researchers, for an effective exchange of ideas and a flowering of future research.

The Banff International Research Station for Mathematical Innovation and Discovery (BIRS) is a collaborative Canada-US-Mexico venture that provides an environment for creative interaction as well as the exchange of ideas, knowledge, and methods within the Mathematical Sciences, with related disciplines and with industry. The research station is located at The Banff Centre in Alberta and is supported by Canada’s Natural Science and Engineering Research Council (NSERC), the U.S. National Science Foundation (NSF), and Alberta’s Advanced Education and Technology.