arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2109.07008cs.LG
收藏

HeMI:异构图中的多视图嵌入

HeMI: Multi-view Embedding in Heterogeneous Graphs

University of Minnesota(明尼苏达大学)

Costas Mavromatis, George Karypis

摘要:

许多现实世界的图涉及不同类型的节点和节点之间的关系,本质上是异构的。异构图的表示学习将异构图丰富的结构和语义嵌入到低维空间中,便于进行节点分类、节点聚类和链接预测等数据采掘任务。在本文中,我们提出了一个自我监督的方法,学习HG表示依赖于不同的HG结构语义(元路径)之间的知识交换和发现。具体来说,通过最大化元路径表示的互信息,我们促进元路径信息的融合和共识,并确保全球共享的语义编码。通过对节点分类、节点聚类和链接预测任务的广泛实验,我们表明,对于所有任务,所提出的自监督方法的性能均优于竞争方法,并将其提高了1%和高达10%。

英文摘要:

Many real-world graphs involve different types of nodes and relations between nodes, being heterogeneous by nature. The representation learning of heterogeneous graphs (HGs) embeds the rich structure and semantics of such graphs into a low-dimensional space and facilitates various data mining tasks, such as node classification, node clustering, and link prediction. In this paper, we propose a self-supervised method that learns HG representations by relying on knowledge exchange and discovery among different HG structural semantics (meta-paths). Specifically, by maximizing the mutual information of meta-path representations, we promote meta-path information fusion and consensus, and ensure that globally shared semantics are encoded. By extensive experiments on node classification, node clustering, and link prediction tasks, we show that the proposed self-supervision both outperforms and improves competing methods by 1% and up to 10% for all tasks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑