Representation Learning on Knowledge Graphs: From Shallow Embedding to Graph Neural Networks

Yizhou Sun (


Knowledge graphs have received tremendous attention recently, due to its wide applications, such as search engines and Q&A systems. Knowledge graph embedding, which aims at representing entities as low-dimensional vectors, and relations as operators on these vectors, has been widely studied and successfully applied to many tasks, such as knowledge reasoning. In this tutorial, we will cover recent representation learning techniques for knowledge graphs, which contains three parts. First, we will review the knowledge graph representation techniques that are usually based on shallow embedding, such as TransE, DisMult, and RotatE. Second, we will discuss the recent progress on how to integrate additional symbolic information, such as logic rules and ontology, for better representation learning on knowledge graphs. In the third part, we will introduce graph neural networks (GNNs) and recent advances on GNNs for heterogeneous information networks, which can be considered as a general form of knowledge graph.