GEMSEC: Graph Embedding with Self Clustering

TL;DR

GEMSEC is a graph embedding algorithm that simultaneously learns node embeddings and clusters, incorporating social network properties to improve community detection quality and efficiency.

Contribution

It introduces a novel method that combines embedding and clustering in a unified framework, extending previous sequence-based graph embedding techniques with social property regularization.

Findings

GEMSEC produces high-quality clusters comparable or superior to existing algorithms.

The method is computationally efficient and robust to hyperparameter variations.

New social network datasets demonstrate the effectiveness of GEMSEC.

Abstract

Modern graph embedding procedures can efficiently process graphs with millions of nodes. In this paper, we propose GEMSEC -- a graph embedding algorithm which learns a clustering of the nodes simultaneously with computing their embedding. GEMSEC is a general extension of earlier work in the domain of sequence-based graph embedding. GEMSEC places nodes in an abstract feature space where the vertex features minimize the negative log-likelihood of preserving sampled vertex neighborhoods, and it incorporates known social network properties through a machine learning regularization. We present two new social network datasets and show that by simultaneously considering the embedding and clustering problems with respect to social properties, GEMSEC extracts high-quality clusters competitive with or superior to other community detection algorithms. In experiments, the method is found to be computationally efficient and robust to the choice of hyperparameters.