Graph Transformer for Communities Detection in Social Networks

Graphs are used in various disciplines such as telecommunication,biological networks,as well as social networks.In large-scale networks,it is challenging to detect the communities by learning the distinct properties of the graph.As deep learning hasmade contributions in a variety of domains,we try to use deep learning techniques to mine the knowledge from large-scale graph networks.In this paper,we aim to provide a strategy for detecting communities using deep autoencoders and obtain generic neural attention to graphs.The advantages of neural attention are widely seen in the field of NLP and computer vision,which has low computational complexity for large-scale graphs.The contributions of the paper are summarized as follows.Firstly,a transformer is utilized to downsample the first-order proximities of the graph into a latent space,which can result in the structural properties and eventually assist in detecting the communities.Secondly,the fine-tuning task is conducted by tuning variant hyperparameters cautiously,which is applied to multiple social networks(Facebook and Twitch).Furthermore,the objective function(crossentropy)is tuned by L0 regularization.Lastly,the reconstructed model forms communities that present the relationship between the groups.The proposed robust model provides good generalization and is applicable to obtaining not only the community structures in social networks but also the node classification.The proposed graph-transformer shows advanced performance on the social networks with the average NMIs of 0.67±0.04,0.198±0.02,0.228±0.02,and 0.68±0.03 on Wikipedia crocodiles,Github Developers,Twitch England,and Facebook Page-Page networks,respectively.

Social Robot Detection Method with Improved Graph Neural Networks

Social robot accounts controlled by artificial intelligence or humans are active in social networks,bringing negative impacts to network security and social life.Existing social robot detection methods based on graph neural networks suffer from the problem of many social network nodes and complex relationships,which makes it difficult to accurately describe the difference between the topological relations of nodes,resulting in low detection accuracy of social robots.This paper proposes a social robot detection method with the use of an improved neural network.First,social relationship subgraphs are constructed by leveraging the user’s social network to disentangle intricate social relationships effectively.Then,a linear modulated graph attention residual network model is devised to extract the node and network topology features of the social relation subgraph,thereby generating comprehensive social relation subgraph features,and the feature-wise linear modulation module of the model can better learn the differences between the nodes.Next,user text content and behavioral gene sequences are extracted to construct social behavioral features combined with the social relationship subgraph features.Finally,social robots can be more accurately identified by combining user behavioral and relationship features.By carrying out experimental studies based on the publicly available datasets TwiBot-20 and Cresci-15,the suggested method’s detection accuracies can achieve 86.73%and 97.86%,respectively.Compared with the existing mainstream approaches,the accuracy of the proposed method is 2.2%and 1.35%higher on the two datasets.The results show that the method proposed in this paper can effectively detect social robots and maintain a healthy ecological environment of social networks.

Spam Short Messages Detection via Mining Social Networks

Short message service （SMS） is now becoming an indispensable way of social communication, and the problem of mobile spam is getting increasingly serious. We propose a novel approach for spare messages detection. Instead of conventional methods that focus on keywords or flow rate filtering, our system is based on mining under a more robust structure： the social network constructed with SMS. Several features, including static features, dynamic features and graph features, are proposed for describing activities of nodes in the network in various ways. Experimental results operated on real dataset prove the validity of our approach.

DFE-GCN: Dual Feature Enhanced Graph Convolutional Network for Controversy Detection

With the development of social media and the prevalence of mobile devices,an increasing number of people tend to use social media platforms to express their opinions and attitudes,leading to many online controversies.These online controversies can severely threaten social stability,making automatic detection of controversies particularly necessary.Most controversy detection methods currently focus on mining features from text semantics and propagation structures.However,these methods have two drawbacks:1)limited ability to capture structural features and failure to learn deeper structural features,and 2)neglecting the influence of topic information and ineffective utilization of topic features.In light of these phenomena,this paper proposes a social media controversy detection method called Dual Feature Enhanced Graph Convolutional Network(DFE-GCN).This method explores structural information at different scales from global and local perspectives to capture deeper structural features,enhancing the expressive power of structural features.Furthermore,to strengthen the influence of topic information,this paper utilizes attention mechanisms to enhance topic features after each graph convolutional layer,effectively using topic information.We validated our method on two different public datasets,and the experimental results demonstrate that our method achieves state-of-the-art performance compared to baseline methods.On the Weibo and Reddit datasets,the accuracy is improved by 5.92%and 3.32%,respectively,and the F1 score is improved by 1.99%and 2.17%,demonstrating the positive impact of enhanced structural features and topic features on controversy detection.

Semi-GSGCN: Social Robot Detection Research with Graph Neural Network

Malicious social robots are the disseminators of malicious information on social networks,which seriously affect information security and network environments.Efficient and reliable classification of social robots is crucial for detecting information manipulation in social networks.Supervised classification based on manual feature extraction has been widely used in social robot detection.However,these methods not only involve the privacy of users but also ignore hidden feature information,especially the graph feature,and the label utilization rate of semi-supervised algorithms is low.Aiming at the problems of shallow feature extraction and low label utilization rate in existing social network robot detection methods,in this paper a robot detection scheme based on weighted network topology is proposed,which introduces an improved network representation learning algorithm to extract the local structure features of the network,and combined with the graph convolution network(GCN)algorithm based on the graph filter,to obtain the global structure features of the network.An end-to-end semi-supervised combination model(Semi-GSGCN)is established to detect malicious social robots.Experiments on a social network dataset(cresci-rtbust-2019)show that the proposed method has high versatility and effectiveness in detecting social robots.In addition,this method has a stronger insight into robots in social networks than other methods.

Real-Time Spammers Detection Based on Metadata Features with Machine Learning

Spammer detection is to identify and block malicious activities performing users.Such users should be identified and terminated from social media to keep the social media process organic and to maintain the integrity of online social spaces.Previous research aimed to find spammers based on hybrid approaches of graph mining,posted content,and metadata,using small and manually labeled datasets.However,such hybrid approaches are unscalable,not robust,particular dataset dependent,and require numerous parameters,complex graphs,and natural language processing(NLP)resources to make decisions,which makes spammer detection impractical for real-time detection.For example,graph mining requires neighbors’information,posted content-based approaches require multiple tweets from user profiles,then NLP resources to make decisions that are not applicable in a real-time environment.To fill the gap,firstly,we propose a REal-time Metadata based Spammer detection(REMS)model based on only metadata features to identify spammers,which takes the least number of parameters and provides adequate results.REMS is a scalable and robust model that uses only 19 metadata features of Twitter users to induce 73.81%F1-Score classification accuracy using a balanced training dataset(50%spam and 50%genuine users).The 19 features are 8 original and 11 derived features from the original features of Twitter users,identified with extensive experiments and analysis.Secondly,we present the largest and most diverse dataset of published research,comprising 211 K spam users and 1 million genuine users.The diversity of the dataset can be measured as it comprises users who posted 2.1 million Tweets on seven topics(100 hashtags)from 6 different geographical locations.The REMS’s superior classification performance with multiple machine and deep learning methods indicates that only metadata features have the potential to identify spammers rather than focusing on volatile posted content and complex graph structures.Dataset and REMS’s codes are available on GitHub(www.github.com/mhadnanali/REMS).

iBole： A Hybrid Multi-Layer Architecture for Doctor Recommendation in Medical Social Networks

In this paper, we try to systematically study how to perform doctor recommendation in medical social net- works （MSNs）. Specifically, employing a real-world medical dataset as the source in our work, we propose iBole, a novel hybrid multi-layer architecture, to solve this problem. First, we mine doctor-patient relationships/ties via a time-constraint probability factor graph model （TPFG）. Second, we extract network features for ranking nodes. Finally, we propose RWR- Model, a doctor recommendation model via the random walk with restart method. Our real-world experiments validate the effectiveness of the proposed methods. Experimental results show that we obtain good accuracy in mining doctor-patient relationships from the network, and the doctor recommendation performance is better than that of the baseline algorithms： traditional Ranking SVM （RSVM） and the individual doctor recommendation model （IDR-Model）. The results of our RWR-Model are more reasonable and satisfactory than those of the baseline approaches.

CAGCN:Centrality-Aware Graph Convolution Network for Anomaly Detection in Industrial Control Systems

In industrial control systems,the utilization of deep learning based methods achieves improvements for anomaly detection.However,most current methods ignore the association of inner components in industrial control systems.In industrial control systems,an anomaly component may affect the neighboring components;therefore,the connective relationship can help us to detect anomalies effectively.In this paper,we propose a centrality-aware graph convolution network(CAGCN)for anomaly detection in industrial control systems.Unlike the traditional graph convolution network(GCN)model,we utilize the concept of centrality to enhance the ability of graph convolution networks to deal with the inner relationship in industrial control systems.Our experiments show that compared with GCN,our CAGCN has a better ability to utilize this relationship between components in industrial control systems.The performances of the model are evaluated on the Secure Water Treatment(SWaT)dataset and the Water Distribution(WADI)dataset,the two most common industrial control systems datasets in the field of industrial anomaly detection.The experimental results show that our CAGCN achieves better results on precision,recall,and F1 score than the state-of-the-art methods.

BPGM: A Big Graph Mining Tool

The design and implementation of a scalable parallel mining system target for big graph analysis has proven to be challenging. In this study, we propose a parallel data mining system for analyzing big graph data generated on a Bulk Synchronous Parallel （BSP） computing model named BSP-based Parallel Graph Mining （BPGM）. This system has four sets of parallel graph mining algorithms programmed in the BSP parallel model and a well-designed workflow engine optimized for cloud computing to invoke these algorithms. Experimental results show that the graph mining algorithm components in BPGM are efficient and have better performance than big cloud-based parallel data miner and BC-BSP.

MBGM: A Graph-Mining Tool Based on MapReduce and BSP

This paper proposes an analytical mining tool for big graph data based on MapReduce and bulk synchronous parallel （BSP） com puting model. The tool is named Mapreduce and BSP based Graphmining tool （MBGM）. The core of this mining system are four sets of parallel graphmining algorithms programmed in the BSP parallel model and one set of data extractiontransformationload ing （ETE） algorithms implemented in MapReduce. To invoke these algorithm sets, we designed a workflow engine which optimized for cloud computing. Finally, a welldesigned data management function enables users to view, delete and input data in the Ha doop distributed file system （HDFS）. Experiments on artificial data show that the components of graphmining algorithm in MBGM are efficient.

Social Opinion Network Analytics in Community Based Customer Churn Prediction

Community based churn prediction,or the assignment of recognising the influence of a customer’s community in churn prediction has become an important concern for firms in many different industries.While churn prediction until recent times have focused only on transactional dataset(targeted approach),the untargeted approach through product advisement,digital marketing and expressions in customer’s opinion on the social media like Twitter,have not been fully harnessed.Although this data source has become an important influencing factor with lasting impact on churn management.Since Social Network Analysis(SNA)has become a blended approach for churn prediction and management in modern era,customers residing online predominantly and collectively decide and determines the momentum of churn prediction,retention and decision support.In existing SNA approaches,customers are classified as churner or non-churner(1 or 0).Oftentimes,the customer’s opinion is also neglected and the network structure of community members are not exploited.Consequently,the pattern and influential abilities of customers’opinion on relative members of the community are not analysed.Thus,the research developed a Churn Service Information Graph(CSIG)to define a quadruple churn category(churner,potential churner,inertia customer,premium customer)for non-opinionated customers via the power of relative affinity around opinionated customers on a direct node to node SNA.The essence is to use data mining technique to investigate the patterns of opinion between people in a network or group.Consequently,every member of the online social network community is dynamically classified into a churn category for an improved targeted customer acquisition,retention and/or decision supports in churn management.

关注社交异配性的社交机器人检测框架

随着社交机器人的迭代,其倾向于与正常用户进行更多交互,对其检测变得更具挑战性。现有检测方法大多基于同配性假设,由于忽视了不同类用户间存在的联系,难以保持良好的检测性能。针对这一问题文章提出一种关注社交异配性的社交机器人检测框架,以社交网络用户间的联系为依据,通过充分挖掘用户社交信息来应对异配影响,并实现更精准的检测。文章分别在同配视角和异配视角下看待用户之间的联系,将社交网络构建为图,通过消息传递机制实现同配边和异配边聚合,以提取节点的频率特征,同时利用图中各节点特征聚合得到社交环境特征,将以上特征混合后用于检测。实验结果表明,文章所提方法在开源数据集上的检测效果优于基线方法,证明了该方法的有效性。

在线社交网络中Spam相册检测方案

提出一种针对Spam相册的检测方案。首先分析了Photo Spam的攻击特点以及与传统Spam的差异,在此基础上构造了12个提取及时且计算高效的特征。利用这些特征提出了有监督学习的检测模型,通过2 356个相册的训练形成Spam相册分类器,实验表明能够正确检测到测试集中100%的Spam相册和98.2%的正常相册。最后将训练后的模型应用到包含315 115个相册的真实数据集中,检测到89 163个Spam相册,正确率达到97.2%。

结合社交网络图的多模态虚假信息检测模型

针对现有虚假信息检测方法主要基于单模态数据分析,检测时忽视了信息之间相关性的问题,提出了结合社交网络图的多模态虚假信息检测模型。该模型使用预训练Transformer模型和图像描述模型分别从多角度提取各模态数据的语义,并通过融合信息传播过程中的社交网络图,在文本和图像模态中加入传播信息的特征,最后使用跨模态注意力机制分配各模态信息权重以进行虚假信息检测。在推特和微博两个真实数据集上进行对比实验,所提模型的虚假信息检测准确率稳定为约88%,高于EANN、PTCA等现有基线模型。实验结果表明所提模型能够有效融合多模态信息,从而提高虚假信息检测的准确率。

基于局部扩展社区发现的学术异常引用群体检测

学术社交网络中的某些学者可能组成异常引用群体,相互之间过度引用彼此的文章以谋取利益。现有的异常群体检测算法大多将社区检测与节点表示学习分离,导致最终异常群体检测性能受限。为此,提出一种基于局部扩展社区发现的异常引用群体检测(GADL)算法。所提算法利用论文研究领域、标题内容等语义信息提取作者异常引用特征;定义基于节点转移相似度、节点社区隶属度、引用异常度和广度优先遍历(BFS)深度的扩展度量函数;结合异常社区发现和异常节点检测,在统一框架下对二者联合优化,可获得最优的异常检测性能。在ACM、DBLP1和DBLP2数据集上,相较于ALP算法,所提算法分别提高了6.07%、5.35%和3.38%。在真实数据集上的实验结果表明,所提算法可有效地检测异常学术引用。

在线社交网络中基于双向动态图注意力网络的异质图谣言检测方法

在线社交网络的普及在帮助人们获取信息的同时,也为谣言的产生和传播创造了条件.现有的基于异质图的谣言检测研究忽略了传播动态性和双向结构特征的重要性.本文提出了一种基于双向动态图注意力网络的异质图谣言检测方法.首先,将谣言传播过程建模为包含消息节点和用户节点的连续时间动态异质图,更细粒度地刻画传播的动态性;其次,使用自注意力和互注意力机制学习不同模态数据之间的隐藏关联,得到包含跨模态信息的表示向量;最后,通过独特的双向图注意力计算方式,在学习时序信息和双向结构信息的同时强化负向关联邻居节点的作用.在真实数据集上的实验结果证明本方案的效果优于其他对比方法.

基于深度强化学习的异常学术引用检测

现有高效识别异常引用的算法存在无法充分利用已知的标签信息或伪标签导致训练过程恶化等问题。为此提出一种融合深度强化学习和图神经网络技术的异常检测方法RACD。异常检测智能体可有效提取作者节点的异常引用特征;异常感知环境建模驱动智能体充分学习已标注数据中的异常特点,发现未标注数据中的潜在异常。通过智能体与环境的不断交互,获得最优的引用异常检测策略。在真实数据集上进行实验,其结果表明,该方法可有效检测异常学术引用。

Rumor detection with self-supervised learning on texts and social graph

Rumor detection has become an emerging and active research field in recent years.At the core is to model the rumor characteristics inherent in rich information,such as propagation patterns in social network and semantic patterns in post content,and differentiate them from the truth.However,existing works on rumor detection fall short in modeling heterogeneous information,either using one single information source only(e.g.,social network,or post content)or ignoring the relations among multiple sources(e.g.,fusing social and content features via simple concatenation).Therefore,they possibly have drawbacks in comprehensively understanding the rumors,and detecting them accurately.In this work,we explore contrastive self-supervised learning on heterogeneous information sources,so as to reveal their relations and characterize rumors better.Technically,we supplement the main supervised task of detection with an auxiliary self-supervised task,which enriches post representations via post self-discrimination.Specifically,given two heterogeneous views of a post(i.e.,representations encoding social patterns and semantic patterns),the discrimination is done by maximizing the mutual information between different views of the same post compared to that of other posts.We devise cluster-wise and instance-wise approaches to generate the views and conduct the discrimination,considering different relations of information sources.We term this framework as self-supervised rumor detection(SRD).Extensive experiments on three real-world datasets validate the effectiveness of SRD for automatic rumor detection on social media.

分享式Spam攻击的轻量级检测方案

Spam攻击是针对社交网络最主要的攻击方式,分享式Spam攻击具有Spam内容的存储与传播分离的新特性,目前没有有效的检测方案。针对这一问题分析了其攻击过程和特征,利用分享式Spam攻击传播和存储的特征设计了轻量级迭代检测算法LIDA,通过目标筛选和内容检测2个步骤实现对分享式Spam的检测。同时,轻量级算法避免了传统算法对每个用户都做深度检测的问题,更具实用性。通过人人网的4次迭代实验,共检测到9 568个Spam账号、30 732个Spam相册以及2 626 780条Spam URL,表明所提的检测算法对于分享式Spam攻击是行之有效的。

基于知识图谱的网络安全漏洞智能检测系统设计

网络安全漏洞智能检测需要依赖大量的真实数据来进行分析,冗余数据与异常数据的存在会导致检测准确性下降;为保障网络系统稳定运行,提出基于知识图谱的网络安全漏洞智能检测系统设计研究;从结构、逻辑模型以及运行模式3个方面设计网络安全漏洞检测器,实现网络安全漏洞智能检测系统硬件设计;系统软件设计通过网络爬虫采集安全漏洞数据,去除冗余数据与异常数据,根据属性信息识别安全漏洞实体,获取安全漏洞属性信息关系,以此为基础,定义安全漏洞知识图谱表示形式,设计安全漏洞知识图谱结构,从而实现安全漏洞知识图谱的构建与可视化;以上述网络设计结果为依据构建网络安全漏洞智能检测整体架构,制定网络安全漏洞智能检测具体流程,从而获取最终网络安全漏洞智能检测结果;实验结果表明,在不同实验工况背景条件下,设计系统应用后的网络安全漏洞漏检率最小值为1.23%,网络安全漏洞检测F1值最大值为9.50,网络安全漏洞检测响应时间最小值为1 ms,证实了设计系统的安全漏洞检测性能更佳。