ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2408.07498
28
2

Wasserstein Gradient Flows of MMD Functionals with Distance Kernel and Cauchy Problems on Quantile Functions

14 August 2024
Richard Duong
Viktor Stein
Robert Beinert
J. Hertrich
Gabriele Steidl
ArXivPDFHTML
Abstract

We give a comprehensive description of Wasserstein gradient flows of maximum mean discrepancy (MMD) functionals Fν:=MMDK2(⋅,ν)\mathcal F_\nu := \text{MMD}_K^2(\cdot, \nu)Fν​:=MMDK2​(⋅,ν) towards given target measures ν\nuν on the real line, where we focus on the negative distance kernel K(x,y):=−∣x−y∣K(x,y) := -|x-y|K(x,y):=−∣x−y∣. In one dimension, the Wasserstein-2 space can be isometrically embedded into the cone C(0,1)⊂L2(0,1)\mathcal C(0,1) \subset L_2(0,1)C(0,1)⊂L2​(0,1) of quantile functions leading to a characterization of Wasserstein gradient flows via the solution of an associated Cauchy problem on L2(0,1)L_2(0,1)L2​(0,1). Based on the construction of an appropriate counterpart of Fν\mathcal F_\nuFν​ on L2(0,1)L_2(0,1)L2​(0,1) and its subdifferential, we provide a solution of the Cauchy problem. For discrete target measures ν\nuν, this results in a piecewise linear solution formula. We prove invariance and smoothing properties of the flow on subsets of C(0,1)\mathcal C(0,1)C(0,1). For certain Fν\mathcal F_\nuFν​-flows this implies that initial point measures instantly become absolutely continuous, and stay so over time. Finally, we illustrate the behavior of the flow by various numerical examples using an implicit Euler scheme and demonstrate differences to the explicit Euler scheme, which is easier to compute, but comes with limited convergence guarantees.

View on arXiv
Comments on this paper