Discrete Latent Graph Generative Modeling with Diffusion Bridges

arXiv:2403.16883v1 Announce Type: new
Abstract: Learning graph generative models over latent spaces has received less attention compared to models that operate on the original data space and has so far demonstrated lacklustre performance. We present GLAD a latent space graph generative model. Unlike most previous latent space graph generative models, GLAD operates on a discrete latent space that preserves to a significant extent the discrete nature of the graph structures making no unnatural assumptions such as latent space continuity. We …

arxiv cs.lg diffusion generative generative modeling graph modeling stat.ml type