GIDN: A Lightweight Graph Inception Diffusion Network for High-efficient Link Prediction. (arXiv:2210.01301v2 [cs.LG] UPDATED)

In this paper, we propose a Graph Inception Diffusion Networks(GIDN) model.
This model generalizes graph diffusion in different feature spaces, and uses
the inception module to avoid the large amount of computations caused by
complex network structures. We evaluate GIDN model on Open Graph Benchmark(OGB)
datasets, reached an 11% higher performance than AGDN on ogbl-collab dataset.

arxiv diffusion graph link prediction network prediction