We give a comprehensive description of Wasserstein gradient flows of maximum mean discrepancy (MMD) functionals towards given target measures on the real line, where we focus on the negative distance kernel . In one dimension, the Wasserstein-2 space can be isometrically embedded into the cone of quantile functions leading to a characterization of Wasserstein gradient flows via the solution of an associated Cauchy problem on . Based on the construction of an appropriate counterpart of on and its subdifferential, we provide a solution of the Cauchy problem. For discrete target measures , this results in a piecewise linear solution formula. We prove invariance and smoothing properties of the flow on subsets of . For certain -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