Deep MMD Gradient Flow without adversarial training

arXiv:2405.06780v1 Announce Type: new
Abstract: We propose a gradient flow procedure for generative modeling by transporting particles from an initial source distribution to a target distribution, where the gradient field on the particles is given by a noise-adaptive Wasserstein Gradient of the Maximum Mean Discrepancy (MMD). The noise-adaptive MMD is trained on data distributions corrupted by increasing levels of noise, obtained via a forward diffusion process, as commonly used in denoising diffusion probabilistic models. The result is a generalization of …

abstract adversarial adversarial training arxiv cs.ai cs.lg data distribution flow generative generative modeling gradient maximum mean modeling noise training type