Neural Flow Diffusion Models: Learnable Forward Process for Improved Diffusion Modelling
This addresses a bottleneck in diffusion modeling for generative AI, offering improved efficiency and versatility, though it is an incremental advancement over existing methods.
The paper tackles the limitations of fixed forward processes in diffusion models, which complicate learning and increase inference costs, by introducing Neural Flow Diffusion Models (NFDM) that support learnable forward processes, achieving state-of-the-art likelihood estimation.
Conventional diffusion models typically relies on a fixed forward process, which implicitly defines complex marginal distributions over latent variables. This can often complicate the reverse process' task in learning generative trajectories, and results in costly inference for diffusion models. To address these limitations, we introduce Neural Flow Diffusion Models (NFDM), a novel framework that enhances diffusion models by supporting a broader range of forward processes beyond the standard Gaussian. We also propose a novel parameterization technique for learning the forward process. Our framework provides an end-to-end, simulation-free optimization objective, effectively minimizing a variational upper bound on the negative log-likelihood. Experimental results demonstrate NFDM's strong performance, evidenced by state-of-the-art likelihood estimation. Furthermore, we investigate NFDM's capacity for learning generative dynamics with specific characteristics, such as deterministic straight lines trajectories, and demonstrate how the framework may be adopted for learning bridges between two distributions. The results underscores NFDM's versatility and its potential for a wide range of applications.