Improving Discriminator Guidance in Diffusion Models
This addresses a specific issue in diffusion model refinement for researchers and practitioners, but it is incremental as it builds on existing guidance methods.
The paper tackles the problem that standard discriminator guidance in diffusion models can increase KL divergence from the real data distribution, and proposes a new training objective that improves sample quality across multiple datasets.
Discriminator Guidance has become a popular method for efficiently refining pre-trained Score-Matching Diffusion models. However, in this paper, we demonstrate that the standard implementation of this technique does not necessarily lead to a distribution closer to the real data distribution. Specifically, we show that training the discriminator using Cross-Entropy loss, as commonly done, can in fact increase the Kullback-Leibler divergence between the model and target distributions, particularly when the discriminator overfits. To address this, we propose a theoretically sound training objective for discriminator guidance that properly minimizes the KL divergence. We analyze its properties and demonstrate empirically across multiple datasets that our proposed method consistently improves over the conventional method by producing samples of higher quality.