Lecture
This lecture by Elsa Cazelles focuses on the streaming computation of optimal weak transport barycenters, covering topics such as Wasserstein distance, barycenter computation, and barycentric projection. The lecture delves into the mathematical formulations and algorithms used to solve the optimal weak transport problem, providing insights into the convergence of proximal algorithms and the challenges in choosing the right transport plan. Practical applications and datasets like MNIST are also discussed, highlighting the significance of weak barycenters in data analysis.