计算机科学

计算机科学 ›› 2025, Vol. 52 ›› Issue (1): 298-306.doi: 10.11896/jsjkx.231100161

• 人工智能 • 上一篇    下一篇

基于多关系图注意力网络的社交机器人检测

孟令君1, 陈鸿昶2, 王庚润2   

  1. 1 郑州大学网络空间安全学院 郑州 450003
    2 信息工程大学国家数字交换系统工程技术研究中心 郑州 450003
  • 收稿日期:2023-11-27 修回日期:2024-05-06 出版日期:2025-01-15 发布日期:2025-01-09
  • 通讯作者: 王庚润(wanggengrun@gmail.com)
  • 作者简介:(893099552@qq.com)
  • 基金资助:
    国家自然科学基金(61803384);嵩山实验室项目(列入河南省科技重大专项)(221100210700-2)

Social Bots Detection Based on Multi-relationship Graph Attention Network

MENG Lingjun1, CHEN Hongchang2, WANG Gengrun2   

  1. 1 School of Cyberspace Security,Zhengzhou University,Zhengzhou 450003,China
    2 National Digital Switching System Engineering & Technological R&D Center,Information Engineering University,Zhengzhou 450003,China
  • Received:2023-11-27 Revised:2024-05-06 Online:2025-01-15 Published:2025-01-09
  • About author:MENG Lingjun,born in 1999,postgra-duate.His main research interests include data analysis,natural language processing and computer vision.
    WANG Gengrun,born in 1987,Ph.D,assistant researcher.His main research interests include telecommunication network security and data processing.
  • Supported by:
    National Natural Science Foundation of China(61803384) and Program of Song Shan Laboratory(included in the management of Major Science and Technology Program of Henan Province)(221100210700-2).

PDF (PC)

179

摘要： 现阶段社交机器人已经广泛存在于社交平台,社交机器人的存在使得网络上的舆论环境可以被人为操纵,这样不仅损害了绿色和谐的网络环境,同时也导致人们正常的网络生活受到极大影响。现有的检测方法可以分为基于特征、基于文本和基于图的方法,其中基于图数据的检测方法大多忽略了图中关系的异质性,并且由于图神经网络存在过渡平滑现象而不能进行深度检测。针对这一问题,提出基于多关系图注意力网络的社交机器人检测方法,在训练时首先将不同关系下的子图抽取出来,然后对子图中的节点采用注意力机制进行聚合,在不同关系下进行节点表示学习并得到节点表示,最后利用通道注意力融合不同关系下的同一节点得到节点表示;同时采用基于LSTM注意力的后连接操作让节点可以自适应地选择邻域进行聚合,以此来缓解过度平滑现象。在Cresci15,Twibot20和MGTAB这3个数据集上的实验结果表明,与11个模型中评价指标的最优值相比,该模型的准确率分别提升了0.47%,1.19%和0.38%,验证了多关系图注意力网络进行社交机器人检测的有效性。

关键词: 异质图, 图注意力, 节点表示学习, LSTM注意力, 社交机器人

Abstract: At present,social bots have gained extensive utilization across social platforms and the existence of social bots makes the public opinion environment on the network artificially manipulated.This not only compromises the integrity of a healthy and harmonious online atmosphere but also significantly disrupts people’s regular online activities.Existing detection methods can be divided into feature-based,text-based,and graph-based methods.However,graph-based detection methods predominantly ignore the heterogeneous relationships,and cannot perform deep detection due to the transition smoothing phenomenon in graph neural networks.To solve the above problems,a social bots detection method based on a multi-relationship graph attention network is proposed.Firstly,we extract subgraphs with different relationships,then apply the attention mechanism to aggregate the nodes within the subgraph and conduct node representation learning across diverse relationships,resulting in the acquisition of node representations.Finally,we use channel attention to fuse the same node under different relationships to obtain node representation,while using the post-connection operation based on LSTM attention to allow nodes to adaptively select neighborhoods for aggregation,thereby alleviating the over-smoothing phenomenon.Experiments are conducted on three datasets:Cresci15,Twibot20,and MGTAB,and the experimental results show that,compared with the optimal values of the evaluation indicators of 11 models,the accuracy of the model is increased by 0.47%,1.19% and 0.38%,respectively,which demonstrates the effectiveness of the multi-relationship graph attention network for social bots detection.

Key words: Heterogeneous graph, Graph attention, Nodes representation learning, LSTM attention, Social bots

中图分类号: 

  • TP391
[1]MARIA K,ILIAS D,ATHENA V.Bot-Detective:An explai-nable Twitter bot detection service with crowdsourcing functionalities[C]//12th International Conference on Management of Digital Ecosystems.New York:Association for Computing Machinery,2020:55-63.
[2]WU Y H,FANG Y Z,SHANG S K,et al.A novel framework for detecting social bots with deep neural networks and active learning[J].European Journal of Medicinal Chemistry:Chimie Therapeutique,2021,211(1):1-16.
[3]ABREU J,GONDIM J,RALHA C.Twitter Bot Detection with Reduced Feature Set[C]//2020 IEEE International Conference on Intelligence and Security Informatics(ISI).USA:Arlington,2020:1-6.
[4]HU F X,LUO W H.Social robot account detection based onmulti-dimensional dynamic feature verification[J].Journal of Foshan University,2023,41(1):23-34.
[5]SNEHA K,EMILIO F.Deep Neural Networks for Bot Detection[J].Information Sciences,2018,467(10):312-322.
[6]GUO Q,XIE H,LI Y,et al.Social Bots Detection via Fusing BERT and Graph Convolutional Networks[J].Symmetry,2022,14(1):1-30.
[7]HAYAWI K,MATHEW S,VENUGOPAL N,et al.DeeProBot:a hybrid deep neural network model for social bot detection based on user profile data[J].Social Network Analysis and Mi-ning,2022,12(1):1-19.
[8]WU J,YE X,MAN Y.BotTriNet:A Unified and Efficient Embedding for Social Bots Detection via Metric Learning[C]//2023 11th International Symposium on Digital Forensics and Security(ISDFS).Turkey:Istanbul,2023:1-6.
[9]WANG X,JI H Y,SHI C,et al.Heterogeneous Graph Attention Network[C]//The World Wide Web Conference(WWW’19).New York:Association for Computing Machinery,2019:2022-2032.
[10]LI Y,JI Y,LI S,et al.Relevance-aware anomalous users’ detection in social network via graph neural network[C]//2021 International Joint Conference on Neural Networks(IJCNN).New York:IEEE,2021:1-8.
[11]FENG S B,WAN H R,WANG N N.SATAR:A Self-supervised Approach to Twitter Account Representation Learning and its Application in Bot Detection[C]//Proceedings of the 30th ACM International Conference on Information & Knowledge Management.New York:Association for Computing Machinery,2021:3808-3817.
[12]FENG S B,WAN H R,WANG N N.BotRGCN:Twitter bot detection with relational graph convolutional networks[C]//Proceedings of the 2021 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining.New York:Association for Computing Machinery,2021:236-239.
[13]YANG Y G,YANG R Y,LI Y Y,et al.RoSGAS:Adaptive Social Bot Detection with Reinforced Self-supervised GNN Architecture Search[J].ACM Transactions on the Web,2023,17(3):1-31.
[14]XU K Y,ZHOU A M,CHEN A L,et al.Social bot detection based on active learning and relational graph convolutional neural networks[J].Journal of Sichuan University(Natural Science Edition),2023,60(5):121-129.
[15]CAI Z J,TAN Z X,LEI Z Y,et al.LMBot:Distilling GraphKnowledge into Language Model for Graph-less Deployment in Twitter Bot Detection[J].arXiv:2306.17408,2023.
[16]SIRUSSTARA J,ALEXANDER N,ALFARISY A,et al.Clickbait Headline Detection in Indonesian News Sites using Robustly Optimized BERT Pre-training Approach(RoBERTa)[C]//2022 3rd International Conference on Artificial Intelligence and Data Sciences(AiDAS).IPOH:Malaysia,2022:1-6.
[17]VASWANI A,SHAZEER N,PARMAR N,et al.Attention IsAll You Need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems(NIPS’17).San Diego:NIPS,2017:6000-6010.
[18]SCHLICHTKRULL M,KIPF T,BLOEM P,et al.Modeling relational data with graph convolutional networks[C]//European Semantic Web Conference.European:Springer,2018:593-607.
[19]HU J,SHEN L,ALBANIE S,et al.Squeeze-and-Excitation Networks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.2017:7132-7141.
[20]LI Q M,HAN Z C,WU X M.Deeper Insights Into Graph Con-volutional Networks for Semi-Supervised Learning[C]//Thirty-Second AAAI Conference on Artificial Intelligence.New York:AAAI Press,2018:3538-3545.
[21]JIANG Y,ZHAO T,CHAI Y,et al.Bidirectional LSTM-CRF models for keyword extraction in Chinese sport news[J].MIPPR 2019:Pattern Recognition and Computer Vision,2020,2(1):11-17.
[22]STEFANO C,ROBERTO D,MARINELLA P,et al.Fame for sale:Efficient detection of fake Twitter followers[J].Decision Support Systems,2015,80(9):56-71.
[23]FENG S B,WAN H,WANG N,et al.TwiBot-20:A Comprehensive Twitter Bot Detection Bench-mark[C]//Proceedings of the 30th ACM International Conference on Information & Knowledge Management,New York:ACM,2020:4485-4494.
[24]SHI S H,QIAO K,CHEN J.MGTAB:A Multi-RelationalGraph-Based Twitter Account Detection Benchmark[C]//IEEE Conference on Computer Vision and Pattern Recognition.New York:IEEE,2023:1-14.
[25]WEI Y L.Classification and regression trees[J].Wiley Interdisciplinary Reviews:Data Mining and Knowledge Discovery,2011,1(1):3-7.
[26]BREIMAN L.Random forests[J].Machine Learning,2004,45:5-32.
[27]SHA A,WANG B,WU X,et al.Semi-Supervised Classification for Hyperspectral Images Using Edge-Conditioned Graph Con-volutional Networks[C]//IEEE International Geoscience and Remote Sensing Symposium(IGARSS 2019).Japan:Yokohama,2019:2690-2693.
[28]PETAR V,GUILLEM C,ARANTXA C,et al.Graph attention networks[C]//International Conference on Learning Representations(ICLR).Washington:ICLR,2018.
[29]WILL H,YING Z T,JURE L.Inductive representation learning on large graphs[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems(NIPS’17).USA:Curran Associates Inc.2017:1025-1035.
[30]HU Z H,DONG Y X,WANG K S,et al.Heterogeneous graph transformer[C]//Proceedings of The Web Conference 2020(WWW’20).USA:Association for Computing Machinery,2020:2704-2710.
[31]YE S,TAN Z,LEI Z,et al.Hofa:Twitter bot detection with homophily-oriented augmentation and frequency adaptive attention[J].arXiv:2306.12870,2023.
[32]SHI S H,QIAO K,YANG J,et al.RF-GNN:Random Forest boosted graph neural network for social bot detection[J].arXiv:2304.08239,2023.
[33]LV Q S,DING M,LIU Q,et al.Are we really making much progress?Revisiting,benchmarking and refining heterogeneous graph neural networks[C]//Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining.New York:ACM,2021:1150-1160.
[34]LU H Y,LIU F,WANG Y B.Social Bot Detection for Dynamic Social Networks Based on Link Prediction[J].Journal of Information Engineering University,2024,25(3):285-291.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发