Semi-Dual Unbalanced Quadratic Optimal Transport: Fast Statistical Rates and Convergent Algorithm
Résumé
In this paper, we derive a semi-dual formulation for the problem of unbalanced quadratic optimal transport and we study its stability properties, namely we give upper and lower bounds for the Bregman divergence of the new objective that hold globally. We observe that the new objective gains even more convexity than in the balanced case. We use this formulation to prove the first results on statistical estimation of UOT potentials and we leverage the extra convexity to recover super-parametric rates. Interestingly, unlike in the balanced case, we do not require the potentials to be smooth. Then, use variable metric descent to solve the semi-dual problem for which we prove convergence at a 1/k rate for strongly convex potentials and exponential convergence in the balanced case when potentials are also smooth. We emphasize that our convergence results have an interest on its own as it generalizes previous convergence results to non-equivalent metrics. Last, we instantiate a proof-of-concept tractable version of our theoretical algorithm that we benchmark on a 2D experiment in the balanced case and on a medium-dimension synthetic experiment in the unbalanced case.
Domaines
Statistiques [math.ST]
Fichier principal
main3.pdf (978.71 Ko)
Télécharger le fichier
Descent_UOTvsSSUOT.png (22.03 Ko)
Télécharger le fichier
FW_vs_Breg_non_smooth.pdf (13.91 Ko)
Télécharger le fichier
descent_ss_balvsunbal.png (209.06 Ko)
Télécharger le fichier
pushforwards.pdf (134.74 Ko)
Télécharger le fichier
rates_2.png (27.37 Ko)
Télécharger le fichier
sinkvssuot.pdf (15.57 Ko)
Télécharger le fichier
train.pdf (36.35 Ko)
Télécharger le fichier
Origine | Fichiers produits par l'(les) auteur(s) |
---|
Format | Figure, Image |
---|
Format | Figure, Image |
---|---|
Origine | Fichiers produits par l'(les) auteur(s) |
Format | Figure, Image |
---|
Format | Figure, Image |
---|
Format | Figure, Image |
---|
Format | Figure, Image |
---|
Format | Figure, Image |
---|