Cometh: A continuous-time discrete-state graph diffusion model
This work addresses graph generation for domains like molecular design, offering incremental improvements in sampling efficiency and quality.
The paper tackled the problem of graph generation by proposing Cometh, a continuous-time discrete-state graph diffusion model, which improved performance over state-of-the-art discrete-state diffusion models, achieving 99.5% on the planar graph dataset and a 12.6% gain on GuacaMol.
Discrete-state denoising diffusion models led to state-of-the-art performance in graph generation, especially in the molecular domain. Recently, they have been transposed to continuous time, allowing more flexibility in the reverse process and a better trade-off between sampling efficiency and quality. Here, to leverage the benefits of both approaches, we propose Cometh, a continuous-time discrete-state graph diffusion model, tailored to the specificities of graph data. In addition, we also successfully replaced the set of structural encodings previously used in the discrete graph diffusion model with a single random-walk-based encoding, providing a simple and principled way to boost the model's expressive power. Empirically, we show that integrating continuous time leads to significant improvements across various metrics over state-of-the-art discrete-state diffusion models on a large set of molecular and non-molecular benchmark datasets. In terms of VUN samples, Cometh obtains a near-perfect performance of 99.5% on the planar graph dataset and outperforms DiGress by 12.6% on the large GuacaMol dataset.