1: \begin{abstract}

2: The aim of this paper is twofold.

3: Based on the geometric Wasserstein tangent space, we first introduce Wasserstein steepest descent flows.

4: These are locally absolutely continuous curves in the Wasserstein space

5: whose tangent vectors point into a steepest descent direction

6: of a given functional.

7: This allows the use of Euler forward schemes instead of minimizing movement schemes 

8: introduced by Jordan, Kinderlehrer and Otto.

9: For locally Lipschitz continuous functionals

10: which are $\lambda$-convex along generalized geodesics, we show that

11: there exists a unique

12: Wasserstein steepest descent flow which coincides with the Wasserstein gradient flow. 

13: 

14: The second aim is to study Wasserstein flows 

15: of the (maximum mean) discrepancy 

16: with respect to certain Riesz kernels. 

17: The crucial part is hereby the treatment of the interaction energy.

18: Although it is not $\lambda$-convex along generalized geodesics,

19: we give analytic expressions for Wasserstein steepest descent flows of

20: the interaction energy starting at Dirac measures.

21: In contrast to smooth kernels, the particle may explode, i.e., the Dirac measure becomes a non-Dirac one.

22: The computation of steepest descent flows amounts to finding equilibrium measures with external fields, which

23: nicely links Wasserstein flows of interaction energies with potential theory.

24: Furthermore, we prove convergence of the minimizing movement scheme 

25: to our Wasserstein steepest descent flow.

26: Finally, we provide analytic  Wasserstein steepest descent flows of discrepancies in one dimension 

27: and numerical simulations in two and three dimensions.

28: 

29: \end{abstract}

30: