LGDC: Latent Graph Diffusion via Spectrum-Preserving Coarsening
By: Nagham Osman , Keyue Jiang , Davide Buffelli and more
Potential Business Impact:
Creates better computer-made pictures of connections.
Graph generation is a critical task across scientific domains. Existing methods fall broadly into two categories: autoregressive models, which iteratively expand graphs, and one-shot models, such as diffusion, which generate the full graph at once. In this work, we provide an analysis of these two paradigms and reveal a key trade-off: autoregressive models stand out in capturing fine-grained local structures, such as degree and clustering properties, whereas one-shot models excel at modeling global patterns, such as spectral distributions. Building on this, we propose LGDC (latent graph diffusion via spectrum-preserving coarsening), a hybrid framework that combines strengths of both approaches. LGDC employs a spectrum-preserving coarsening-decoarsening to bidirectionally map between graphs and a latent space, where diffusion efficiently generates latent graphs before expansion restores detail. This design captures both local and global properties with improved efficiency. Empirically, LGDC matches autoregressive models on locally structured datasets (Tree) and diffusion models on globally structured ones (Planar, Community-20), validating the benefits of hybrid generation.
Similar Papers
Decoupling Complexity from Scale in Latent Diffusion Model
CV and Pattern Recognition
Makes pictures and videos with any detail.
Graph-Aware Diffusion for Signal Generation
Machine Learning (CS)
Creates realistic data for networks and recommendations.
H-LDM: Hierarchical Latent Diffusion Models for Controllable and Interpretable PCG Synthesis from Clinical Metadata
Machine Learning (CS)
Creates realistic heart sounds for training doctors.