Discriminator Guidance for Autoregressive Diffusion Models
This work addresses the challenge of enhancing generative modeling for discrete data, such as molecular graphs, by adapting discriminator guidance from continuous to autoregressive diffusion models, representing an incremental advancement in the field.
The authors tackled the problem of improving generative performance in autoregressive diffusion models by introducing discriminator guidance, showing that it enables exact sampling from the data distribution with an optimal discriminator and improves molecular graph generation with a sequential Monte Carlo algorithm for sub-optimal cases.
We introduce discriminator guidance in the setting of Autoregressive Diffusion Models. The use of a discriminator to guide a diffusion process has previously been used for continuous diffusion models, and in this work we derive ways of using a discriminator together with a pretrained generative model in the discrete case. First, we show that using an optimal discriminator will correct the pretrained model and enable exact sampling from the underlying data distribution. Second, to account for the realistic scenario of using a sub-optimal discriminator, we derive a sequential Monte Carlo algorithm which iteratively takes the predictions from the discriminator into account during the generation process. We test these approaches on the task of generating molecular graphs and show how the discriminator improves the generative performance over using only the pretrained model.