Published: Jul 14, 2024

Graph diffusion convolution is a geometric deep learning architecture that aggregates information from higher-order network neighbors through a generalized graph diffusion to enhance model robustness to noisy and incomplete datasets. Paper at ICML.