SubRank: Subgraph Embeddings via a Subgraph Proximity Measure

Representation learning for graph data has gained a lot of attention in recent years. However, state-of-the-art research is focused mostly on node embeddings, with little effort dedicated to the closely related task of computing subgraph embeddings. Subgraph embeddings have many applications, such as community detection, cascade prediction, and question answering. In this work, we propose a subgraph to subgraph proximity measure as a building block for a subgraph embedding framework. Experiments on real-world datasets show that our approach, SubRank, outperforms state-of-the-art methods on several important data mining tasks.