Simple Guidance Mechanisms for Discrete Diffusion Models
This work addresses the problem of controllable generation for discrete data in fields such as genomics and chemistry, representing an incremental improvement over existing methods.
The paper tackles the challenge of controllable generation for discrete data by deriving classifier-free and classifier-based guidance mechanisms for discrete diffusion models, and introduces a uniform noise diffusion approach with a continuous-time variational lower bound, achieving state-of-the-art performance in domains like genomic sequences and small molecule design.
Diffusion models for continuous data gained widespread adoption owing to their high quality generation and control mechanisms. However, controllable diffusion on discrete data faces challenges given that continuous guidance methods do not directly apply to discrete diffusion. Here, we provide a straightforward derivation of classifier-free and classifier-based guidance for discrete diffusion, as well as a new class of diffusion models that leverage uniform noise and that are more guidable because they can continuously edit their outputs. We improve the quality of these models with a novel continuous-time variational lower bound that yields state-of-the-art performance, especially in settings involving guidance or fast generation. Empirically, we demonstrate that our guidance mechanisms combined with uniform noise diffusion improve controllable generation relative to autoregressive and diffusion baselines on several discrete data domains, including genomic sequences, small molecule design, and discretized image generation.