Global and Local Topology-Aware Graph Generation via Dual Conditioning Diffusion

Graph generation plays an important role in various domains such as molecular design, protein prediction, and drug discovery. However, generating graphstructured data poses challenges due to the complex dependencies inherent in graphs, spanning from intricate local substructures to broad global topologies. Although recent advances in graph-generative models have made notable progress, traditional node-level generative paradigms may have difficulty simultaneously capturing the multiscale dependencies in graphs. To address these challenges, we propose a unified latent diffusion model that jointly learns local and global topological information, enabling effective and efficient graph generation. Besides, our approach introduces a dual conditioning mechanism designed to promote dynamic interaction between local and global information, equipping the generative model with global and local awareness to better capture the coupled dependencies within graphs. Our method can largely promote the joint modeling of global and local information and substantially improve the quality of the generated graphs. Extensive experiments consistently demonstrate the effectiveness of our method. Recently, some effective models have incorporated global information to enhance the graph generation process. For instance, SubgDiff (Zhang et al., 2024a) introduces a subgraph prediction module that integrates substructure information into diffusion models, thereby improving molecular representation learning. Similarly, Graphusion (Yang et al., 2024) leverages graph-level pseudo-labels derived from clustering algorithms to provide informative guidance during the generation process.