Sobolev regularization on the witness function enables global convergence of MMD gradient flows for both sampling and generative modeling without isoperimetric assumptions.
Journal of Machine Learning Research , volume=
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3verdicts
UNVERDICTED 3representative citing papers
Geometric tempering yields exponential convergence bounds for both Wasserstein and Fisher-Rao flows but produces no speedup in the Fisher-Rao metric, with new adaptive schedules derived from the tempered dynamics.
MIND uses sliced Wasserstein distance on Inception features to evaluate generative models, matching FID performance with 10x fewer samples and 100x faster computation while being more robust to moment-matching attacks.
citing papers explorer
-
Sobolev Regularized MMD Gradient Flow
Sobolev regularization on the witness function enables global convergence of MMD gradient flows for both sampling and generative modeling without isoperimetric assumptions.
-
Properties and limitations of geometric tempering for gradient flow dynamics
Geometric tempering yields exponential convergence bounds for both Wasserstein and Fisher-Rao flows but produces no speedup in the Fisher-Rao metric, with new adaptive schedules derived from the tempered dynamics.
-
MIND: Monge Inception Distance for Generative Models Evaluation
MIND uses sliced Wasserstein distance on Inception features to evaluate generative models, matching FID performance with 10x fewer samples and 100x faster computation while being more robust to moment-matching attacks.