Learning the Implicit Semantic Representation on Graph-Structured Data

Existing representation learning methods in graph convolutional networks are mainly designed by describing the neighborhood of each node as a perceptual whole, while implicit semantic associations behind highly complex interactions graphs largely unexploited. In this paper, we propose Semantic Graph Convolutional Networks (SGCN) that explores semantics latent semantic-paths graphs. previous work, there explorations via meta-paths. However, these rely on explicit heterogeneous information is hard to be obtained large amount graph-structured data. SGCN first breaks through restriction leveraging dynamically and automatically during aggregating process. To evaluate our idea, conduct sufficient experiments several standard datasets, empirical results show superior performance model (Our code available online at https://github.com/WLiK/SGCN_SemanticGCN ).