CCVFM replaces the inner noise source in hierarchical rectified flow matching with a data-informed Gaussian mixture surrogate from a Sinkhorn coreset, yielding a closed-form conditional velocity law and competitive few-step generation on MNIST, CIFAR-10, ImageNet-32, and CelebA-HQ.
GANs trained by a two time-scale update rule converge to a local Nash equilibrium
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
method 1
citation-polarity summary
fields
stat.ML 1years
2026 1verdicts
UNVERDICTED 1roles
method 1polarities
use method 1representative citing papers
citing papers explorer
-
Coreset-Induced Conditional Velocity Flow Matching
CCVFM replaces the inner noise source in hierarchical rectified flow matching with a data-informed Gaussian mixture surrogate from a Sinkhorn coreset, yielding a closed-form conditional velocity law and competitive few-step generation on MNIST, CIFAR-10, ImageNet-32, and CelebA-HQ.