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.

翻译：本文提出了一种图初始扩散网络（Graph Inception Diffusion Networks，简称GIDN）模型。该模型在不同特征空间中泛化图扩散，并利用初始模块（inception module）避免复杂网络结构带来的大量计算。我们在开放图基准（Open Graph Benchmark，简称OGB）数据集上对GIDN模型进行评估，在ogbl-collab数据集上达到了比AGDN模型高11%的性能。