Unifying GANs and Score-Based Diffusion as Generative Particle Models
This work addresses the theoretical gap between different generative modeling paradigms for researchers in machine learning, though it appears incremental as it builds on existing particle and adversarial methods.
The paper tackles the perceived opposition between particle-based generative models (like diffusion models) and GANs by proposing a unifying framework that frames generator training as a generalization of particle models, showing that generators are optional and enabling novel hybrid models like generator-augmented diffusion and generator-free GANs.
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.