19 – 25 November 2023, Oberwolfach Mathematisches Forschunginstitute.
Variational and stochastic flows are now ubiquitous in machine learning and generative modeling. Indeed, many such models can be interpreted as flows from a latent distribution to the sample distribution and training corresponds to finding the right flow vector field. Optimal transport and diffeomorphic flows provide powerful frameworks to analyze such trajectories of distributions with elegant notions from differential geometry, such as geodesics, gradient and Hamiltonian flows. Recently, mean field control and mean field games offer a general optimal control variational problems on the learning problem. How do these tools lead us to a better understanding and further development of machine learning and generative models? The Oberwolfach Seminar will address the topic from different points of view taking in particular recent developments in machine learning into account. The target audience is PhD students and post-doctoral researchers wishing to be quickly immersed in this modern, active research area. Priority will be given to young, motivated researchers.
Official webpage of the seminar.
Lectures were given by (chronological order):
- Bernhard Schmitzer (Göttingen)
- Gabriele Steidl (T.U. Berlin)
- Wuchen Li (USC Chapel Hill)
- F.-X. Vialard (Some slides associated with the lecture)
