Implement tuning-free Wasserstein gradient flows via the FUSE adaptive step size of Sharrock and Nemeth (2025). Example implementation in JAX available here.
The squared distance travelled (d_s^2) and the gradient energy (g_s^2) could be stored as diagnostics and relate to #4 .
Implement tuning-free Wasserstein gradient flows via the FUSE adaptive step size of Sharrock and Nemeth (2025). Example implementation in JAX available here.
The squared distance travelled (
d_s^2) and the gradient energy (g_s^2) could be stored as diagnostics and relate to #4 .