Enhancing graph link prediction with sequential latent variable diffusion models

Wenchuan Zhang

doi:10.1016/j.patcog.2026.114957

Enhancing graph link prediction with sequential latent variable diffusion models

2026-09-23

0

PRE

AI

DOI：10.1016/j.patcog.2026.114957

Original

Original request for help

Save

Abstract

En 中文

• Unveiled limitations of traditional method in link prediction. • Introduced multiple, independent latent variables for enhanced expressivity. • Developed a Markov chain-based method for link prediction tasks. • Applied diffusion model principles to link prediction. • Demonstrated improved prediction accuracy on datasets and detailed visualization.

Enhancing graph link prediction with sequential latent variable diffusion models

Abstract

Journal

Organization

Cited Papers