Unifying GANs and Score-Based Diffusion as Generative Particle Models
Résumé
Particle-based deep generative models, such as gradient flows and score-based diffusion models, have recently gained traction thanks to their striking performance. Their principle of displacing particle distributions using differential equations is conventionally seen as opposed to the previously widespread generative adversarial networks (GANs), which involve training a pushforward generator network. In this paper we challenge this interpretation, and propose a novel framework that unifies particle and adversarial generative models by framing generator training as a generalization of particle models. This suggests that a generator is an optional addition to any such generative model. Consequently, integrating a generator into a score-based diffusion model and training a GAN without a generator naturally emerge from our framework. We empirically test the viability of these original models as proofs of concepts of potential applications of our framework.
Origine | Fichiers produits par l'(les) auteur(s) |
---|---|
Licence |
![]()
Décrit hal-04107806 Objet présenté à une conférence Jean-Yves Franceschi, Mike Gartrell, Ludovic Dos Santos, Thibaut Issenhuth, Emmanuel de Bézenac, et al.. Unifying GANs and Score-Based Diffusion as Generative Particle Models. Thirty-seventh Conference on Neural Information Processing Systems, Neural Information Processing Systems Foundation, Dec 2023, New Orleans, LA, United States. ⟨hal-04107806v3⟩