Reduce, Reuse, Recycle: Compositional Generation with Energy-Based Diffusion Models and MCMC
This work addresses compositional generation issues in diffusion models, which is an incremental advancement for generative modeling in domains like image synthesis.
The paper tackled the problem of compositional generation failures in diffusion models by identifying the sampler as the cause and proposing new MCMC-inspired samplers and an energy-based parameterization, resulting in notable improvements across tasks like ImageNet modeling and text-to-image generation.
Since their introduction, diffusion models have quickly become the prevailing approach to generative modeling in many domains. They can be interpreted as learning the gradients of a time-varying sequence of log-probability density functions. This interpretation has motivated classifier-based and classifier-free guidance as methods for post-hoc control of diffusion models. In this work, we build upon these ideas using the score-based interpretation of diffusion models, and explore alternative ways to condition, modify, and reuse diffusion models for tasks involving compositional generation and guidance. In particular, we investigate why certain types of composition fail using current techniques and present a number of solutions. We conclude that the sampler (not the model) is responsible for this failure and propose new samplers, inspired by MCMC, which enable successful compositional generation. Further, we propose an energy-based parameterization of diffusion models which enables the use of new compositional operators and more sophisticated, Metropolis-corrected samplers. Intriguingly we find these samplers lead to notable improvements in compositional generation across a wide set of problems such as classifier-guided ImageNet modeling and compositional text-to-image generation.