<ul data-eligibleForWebStory="true"><li>Autoregressive models in the transformer ecosystem provide efficiency, scalability, and seamless workflows but require explicit sequence order which conflicts with unordered graphs.</li><li>Diffusion models maintain permutation invariance and allow one-shot generation but necessitate numerous denoising steps, additional features, and high computational costs.</li><li>MAG is a novel diffusion-free graph generation framework inspired by visual autoregressive methods, based on next-scale prediction.</li><li>MAG utilizes a hierarchy of latent representations to generate graph scales progressively without explicit node ordering.</li><li>Experiments on generic and molecular graph datasets showed MAG's potential, achieving speedups up to three orders of magnitude over existing methods while maintaining high-quality generation.</li></ul>

Diffusion-Free Graph Generation with Next-Scale Prediction

Discover more