Why does the author here suggest that we look for absolutely continuous curves as solutions to the following differential inclusion (or LESS generally a differential equation)?
Here the $\partial F$ is the subgradient. The source is Wasserstein Gradient Flows by Santambrogio.