High-performance community detection in social networks using a deep transitive autoencoder

A Comprehensive Survey on Community Detection With Deep Learning

Detecting a community in a network is a matter of discerning the distinct features and connections of a group of members that are different from those in other communities. The ability to do this is of great significance in network analysis. However, beyond the classic spectral clustering and statistical inference methods, there have been significant developments with deep learning techniques for community detection in recent years-particularly when it comes to handling high-dimensional network data. Hence, a comprehensive review of the latest progress in community detection through deep learning is timely. To frame the survey, we have devised a new taxonomy covering different state-of-the-art methods, including deep learning models based on deep neural networks (DNNs), deep nonnegative matrix factorization, and deep sparse filtering. The main category, i.e., DNNs, is further divided into convolutional networks, graph attention networks, generative adversarial networks, and autoencoders. The popular benchmark datasets, evaluation metrics, and open-source implementations to address experimentation settings are also summarized. This is followed by a discussion on the practical applications of community detection in various domains. The survey concludes with suggestions of challenging topics that would make for fruitful future research directions in this fast-growing deep learning field.

Robust, scalable, and informative clustering for diverse biological networks

Clustering molecular data into informative groups is a primary step in extracting robust conclusions from big data. However, due to foundational issues in how they are defined and detected, such clusters are not always reliable, leading to unstable conclusions. We compare popular clustering algorithms across thousands of synthetic and real biological datasets, including a new consensus clustering algorithm-SpeakEasy2: Champagne. These tests identify trends in performance, show no single method is universally optimal, and allow us to examine factors behind variation in performance. Multiple metrics indicate SpeakEasy2 generally provides robust, scalable, and informative clusters for a range of applications.

Gravity-Law Based Critical Bots Identification in Large-Scale Heterogeneous Bot Infection Network

The explosive growth of botnets has posed an unprecedented potent threat to the internet. It calls for more efficient ways to screen influential bots, and thus precisely bring the whole botnet down beforehand. In this paper, we propose a gravity-based critical bots identification scheme to assess the influence of bots in a large-scale botnet infection. Specifically, we first model the propagation of the botnet as a Heterogeneous Bot Infection Network (HBIN). An improved SEIR model is embedded into HBIN to extract both heterogeneous spatial and temporal dependencies. Within built-up HBIN, we elaborate a gravity-based influential bots identification algorithm where intrinsic influence and infection diffusion influence are specifically designed to disclose significant bots traits. Experimental results based on large-scale sample collections from the implemented prototype system demonstrate the promising performance of our scheme, comparing it with other state-of-the-art baselines.

Identification of rumour stances by considering network topology and social media comments

Online social media (OSM) has become a hotbed for the rapid dissemination of disinformation or faked news. In order to track and limit the spread of faked news, we study stance identification of comments posted on OSM, where the stance can denote the comment’s semantics. In this article, we propose a framework for identification of rumour stances, combining network topology and OSM comments. We construct a vector matrix of comments and words via OTI (optimisation term frequency–inverse document frequency). To better identify the stances, we introduce another vector matrix with novel or special attribute, that is, network topology among the users. Variant autoencoder (VAE) is then applied for dimensionality reduction and optimisation of these vector matrices which are then combined into an integrated matrix [Formula: see text], tempered by two parameters [Formula: see text] and [Formula: see text]. Finally, the matrix is fed into a neural network for final rumour stance identification. Experimental evaluations show that our proposed approach outperforms some state-of-the-art methods and achieves a high precision of 90.26% and F1-score of 88.58%.