可扩展图学习的分层 Transformer（英文版）

英文标题：Hierarchical Transformer for Scalable Graph Learning中文摘要：本文介绍了一种名为 HSGT 的层级可伸缩图变换器，通过利用层次结构和基于采样的训练方法，高效地更新和存储多尺度信息，并使用 Transformer 块有效地捕获和聚合多级信息以实现节点表征学习任务在大规模图上的高性能表现。英文摘要：Graph Transformer is gaining increasing attention in the field of machinelearning and has demonstrated state-of-the-art performance on benchmarks forgraph representation learning. However, as current implementations of GraphTransformer primarily focus on learning representations of small-scale graphs,the quadratic compl