Deep MMD Gradient Flow without adversarial training.

Alexandre Galashov, Valentin De Bortoli, Arthur Gretton

Browse the full ICLR paper archive.