Graph Partition Neural Networks for Semi-Supervised Classification
Renjie Liao University of Toronto Uber ATG Toronto Vector Institute、Marc Brockschmidt Microsoft Research Cambridge、Daniel Tarlow∗ Google Brain \ANDAlexander L. Gaunt Microsoft Research Cambridge、Raquel Urtasun University of Toronto Uber ATG Toronto Vector Institute、Richard Zemel University of Toronto Vector Institute Canadian Institute for Advanced Research Work done partially while author was at Microsoft Research.
Abstract
We present graph partition neural networks (GPNN), an extension of graph neural networks (GNNs) able to handle extremely large graphs. GPNNs alternate between locally propagating information between nodes in small subgraphs and globally propagating information between the subgraphs. To efficiently partition graphs, we experiment with several partitioning algorithms and also propose a novel variant for fast processing of large scale graphs. We extensively test our model on a variety of semi-supervised node classification tasks. Experimental results indicate that GPNNs are either superior or comparable to state-of-the-art methods on a wide variety of datasets for graph-based semi-supervised classification. We also show that GPNNs can achieve similar performance as standard GNNs with fewer propagation steps.
中文速览
大规模图数据上的信息传播效率低、内存消耗大，是图神经网络（Graph Neural Networks, GNNs）长期面临的瓶颈。作者提出了图划分神经网络（Graph Partition Neural Networks, GPNNs），核心思路是先用图划分算法把大图切成若干互不重叠的子图，再交替执行"子图内部同步传播"和"子图之间跨界传播"两个阶段，从而在保持并行计算优势的同时，大幅减少整体所需的消息数量和计算深度。在多个半监督节点分类基准数据集上的实验表明，GPNN 的性能与当前最优方法持平甚至更好，且能用更少的传播步骤达到与标准 GNN 相当的效果。这项工作为把图神经网络扩展到超大规模图提供了一条兼顾计算效率和模型表达能力的可行路径。
原文 arXiv:1803.06272;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1803.06272v1