$h$-Wasserstein barycenters
Camilla Brizzi, Gero Friesecke, Tobias Ried

TL;DR
This paper extends the concept of Wasserstein barycenters to general convex costs, proving uniqueness and Monge form of optimal plans using a novel injectivity approach.
Contribution
It generalizes Wasserstein barycenters beyond quadratic costs, establishing equivalence of formulations and proving uniqueness and Monge structure with a new injectivity method.
Findings
Multi-marginal optimal plan is unique.
Optimal plan is of Monge form.
Introduces a new injectivity-based approach.
Abstract
We generalize the notion and theory of Wasserstein barycenters introduced by Agueh and Carlier (2011) from the quadratic cost to general smooth strictly convex costs with non-degenerate Hessian. We show the equivalence between a coupled two-marginal and a multi-marginal formulation and establish that the multi-marginal optimal plan is unique and of Monge form. To establish the latter result we introduce a new approach which is not based on explicitly solving the optimality system, but instead deriving a quantitative injectivity estimate for the (highly non-injective) map from -point configurations to their -barycenter on the support of an optimal multi-marginal plan.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsGeometric Analysis and Curvature Flows · Point processes and geometric inequalities
