KI Seminar (online)

The Kantorovich Initiative Seminar takes place online. To attend, please register for the KI mailing list.

 

From Diffusion Models to Schrödinger Bridges - When Generative Modeling meets Optimal Transport
Arnaud Doucet (University of Oxford/Google Deepmind)
Denoising Diffusion models have revolutionized generative modeling. Conceptually, these methods define a transport mechanism from a …
2025, Apr 10 10:00 AM PDT Online (zoom)
Video
Metric Wasserstein extrapolation, semi-discrete convex order and Laguerre tessellation fitting
Andrea Natale (Researcher at Inria Université de Lille)
In this talk, we study a variational problem that provides a way to extend minimizing geodesics connecting two given probability …
2025, Apr 9 11:00 AM PDT UBC - ESB4133 (PIMS) & Online (zoom)
Gluing methods for quantitative stability of optimal transport maps
Quentin Merigot (Université Paris-Saclay)

Stability results have important applications in statistics (including on the estimation of OT maps from random data, and the matching …

2025, Mar 6 10:00 AM PST Online (zoom)
Statistics for entropic optimal transport with decreasing regularisation
Gilles Mordant (Göttingen)
In this talk, we will discuss the question of establishing CLTs for empirical entropic optimal transport when choosing the …
2025, Feb 20 10:00 AM PST Online (zoom)
Slides Video
Transport- and Measure-Theoretic Approaches for Modeling, Identifying, and Forecasting Dynamical Systems
Yunan Yang (Cornell University)
Measures provide valuable insights into long-term and global behaviors across a broad range of dynamical systems. In this talk, we …
2025, Jan 23 10:00 AM PST Online (zoom)
Slides Video
Kernel Approximation of Wasserstein and Fisher-Rao Gradient flows
Jia-Jie Zhu (Weierstrass Institute for Applied Analysis and Stochastics, Berlin)
Gradient flows have emerged as a powerful framework for analyzing machine learning and statistical inference algorithms. Motivated by …
2024, Dec 16 10:45 AM PST Online (zoom)
Slides Video
Distributional Preference Alignment of Large Language Models via Optimal Transport
Yousef Mroueh (IBM)

Current LLM alignment techniques use pairwise human preferences at a sample level, and as such, they do not imply an alignment on the …

2024, Dec 13 9:00 AM PST ESB4133 and Online (zoom)
Slides Video
On Parameterizing Optimal Transport with Elastic Costs
Marco Cuturi (Apple ML Research, Paris)
I will introduce our recent work on parameterising OT problems with elastic costs, i.e. ground costs that mix the classic squared …
2024, Dec 9 2:45 PM PST Online (zoom)
Quadratically regularized OT
Alberto González-Sanz (Columbia University)
The quadratically regularized optimal transport problem (QOT) has emerged in the literature as a sparse alternative to entropic …
2024, Nov 21 10:00 AM PST Online (zoom)
Slides Video
Gradient flows in the geometry of Sinkhorn divergences
Hugo Lavenant (Bocconi University)
What happens to Wasserstein gradient flows if one uses entropic optimal transport instead of classical optimal transport? I will …
2024, Oct 31 10:00 AM PDT ESB4133 and Online (zoom)
Slides Video