Neural Graph Generator: Feature-Conditioned Graph Generation using Latent Diffusion Models

TL;DR

This work introduces the Neural Graph Generator (NGG), a novel approach which utilizes conditioned latent diffusion models for graph generation, demonstrating a remarkable capacity to model complex graph patterns, offering control over the graph generation process.

Abstract

Graph generation has emerged as a crucial task in machine learning, with significant challenges in generating graphs that accurately reflect specific properties. Existing methods often fall short in efficiently addressing this need as they struggle with the high-dimensional complexity and varied nature of graph properties. In this paper, we introduce the Neural Graph Generator (NGG), a novel approach which utilizes conditioned latent diffusion models for graph generation. NGG demonstrates a remarkable capacity to model complex graph patterns, offering control over the graph generation process. NGG employs a variational graph autoencoder for graph compression and a diffusion process in the latent vector space, guided by vectors summarizing graph statistics. We demonstrate NGG's versatility across various graph generation tasks, showing its capability to capture desired graph properties and generalize to unseen graphs. We also compare our generator to the graph generation capabilities of different LLMs. This work signifies a shift in graph generation methodologies, offering a more practical and efficient solution for generating diverse graphs with specific characteristics.

Figures (4)

• Figure 1: Overview of the proposed architecture. The variational graph autoencoder is responsible for generating a compressed latent representation $\boldsymbol{z}$ for each graph $\boldsymbol{G}$. Those representations are fed to the diffusion model which operates in that latent space adding noise to $\boldsymbol{z}$ resulting to $\boldsymbol{z}_T$ . The denoising process is conditioned on the encoding (output of $\tau_\theta$) of the vector that contains the graph's properties. The output of the diffusion model is passed on to the decoder which generates a graph.
• Figure 2: Example of two graphs generated by the proposed NGG model given condition vectors $\mathbf{c}_1$ and $\mathbf{c}_2$.
• Figure 3: Prompting examples for LLM graph generation
• Figure 4: Importance of graph properties, given the number of nodes, edges, and one of the remaining 13 properties as input.