CommunityGAN: Community Detection with Generative Adversarial Nets

Abstract

Community detection refers to the task of discovering groups of vertices sharing similar properties or functions so as to understand the network data. With the recent development of deep learning, graph representation learning techniques are also utilized for community detection. However, the communities can only be inferred by applying clustering algorithms based on learned vertex embeddings. These general cluster algorithms like K-means and Gaussian Mixture Model cannot output much overlapped communities, which have been proved to be very common in many real-world networks. In this paper, we propose CommunityGAN, a novel community detection framework that jointly solves overlapping community detection and graph representation learning. First, unlike the embedding of conventional graph representation learning algorithms where the vector entry values have no specific meanings, the embedding of CommunityGAN indicates the membership strength of vertices to communities. Second, a specifically designed Generative Adversarial Net (GAN) is adopted to optimize such embedding. Through the minimax competition between the motif-level generator and discriminator, both of them can alternatively and iteratively boost their performance and finally output a better community structure. Extensive experiments on synthetic data and real-world tasks demonstrate that CommunityGAN achieves substantial community detection performance gains over the state-of-the-art methods.

DMFO-CD: A Discrete Moth-Flame Optimization Algorithm for Community Detection

“…The proposed DMFO-CD algorithm was tested on eleven real-world network datasets known as social networks consists of Zachary's karate club (karate) [92], American College football (football) [2], Bottlenose Dolphins (dolphins) [93], co-purchase political books (polbooks) [94], WebKB network [95] includes four datasets, adjective Noun network (adjnoun) [12], Email-Eu-core network (email-Eu) [96], and DBLP network (dblp) [97]. The details and statistical information of the networks are described as follows.…”

A Survey on the Recent Advances of Deep Community Detection

“…CommunityGAN [68] is a community detection scheme that jointly solves overlapping community detection and graph representation learning. The CommunityGAN strategy learns network embeddings like AGM (Affiliation Graph Model) through a specifically designed GAN.…”

“…An idea is to combine the effective GANS with affiliation graph models (AGM), which can model densely overlapping community structures. Thus, the performance of GANs and the direct vertex-community membership representation of AGMs join forces to solve the dense overlapping issues [68]. In addition, the scheme presented in [69] is working on overlaps.…”

Learning Concept Interestingness for Identifying Key Structures From Social Networks

“…Compared with traditional community detection approaches, deep learning-based methods aim to identify community structures by creating more powerful representations of node attributes and community structures [27]. Concretely, depending on the used learning strategies, deep learning-based methods for finite and infinite community detection fall into five main categories: convolutional neural network (CNN)based [28], auto-encoder-based [29], generative adversarial network-based [30], graph embedding-based [31] [32] and graph neural network (GNN)-based [33] [34]. Comprehensive surveys [27] [35] of community detection approaches are referred.…”