Generating graphs with fidelity and diversity using SteinGen.
DeFoG, a novel graph generation framework based on discrete flow matching (DFM), outperforms diffusion models in terms of efficiency and flexibility by decoupling the training and sampling stages and introducing algorithmic improvements for both.
Graph Beta Diffusion (GBD) is a novel generative model that leverages the flexibility of beta distributions to effectively capture the diverse statistical characteristics of graph data, including discrete structures and continuous node attributes, leading to improved realism in generated graphs.
GraphMaker, a novel diffusion model, effectively generates large attributed graphs by asynchronously denoising node attributes and graph structure, addressing scalability challenges and demonstrating superior performance in preserving data utility for machine learning model development and benchmarking.