计算机科学

计算机科学 ›› 2023, Vol. 50 ›› Issue (11): 49-54.doi: 10.11896/jsjkx.221000043

• 数据库&大数据&数据科学 • 上一篇    下一篇

基于边推断增强对比学习的社交媒体谣言检测模型

刘楠, 张凤荔, 尹嘉奇, 陈学勤, 王瑞锦   

  1. 电子科技大学信息与软件工程学院 成都 610054
  • 收稿日期:2022-10-09 修回日期:2022-12-30 出版日期:2023-11-15 发布日期:2023-11-06
  • 通讯作者: 尹嘉奇(yinjiaqi@Gmail.com)
  • 作者简介:(ln@std.uestc.edu.cn)
  • 基金资助:
    国家自然科学基金(62271128);四川省区域创新合作项目(2020YFQ0018);四川省科技计划重点研发项目(2022ZDZX0004,2023YFG0029,2023YFG0150,2023ZHCG004,2022YFG0212,2021YFS0391,2021YFG0027)

Rumor Detection Model on Social Media Based on Contrastive Learning with Edge-inferenceAugmentation

LIU Nan, ZHANG Fengli, YIN Jiaqi, CHEN Xueqin, WANG Ruijin   

  1. School of Information and Software Engineering,University of Electronic Science and Technology of China,Chengdu 610054,China
  • Received:2022-10-09 Revised:2022-12-30 Online:2023-11-15 Published:2023-11-06
  • About author:LIU Nan,born in 1995.Ph.D.Her main research interests include social network analysis and deep representation learning.YIN Jiaqi,born in 1986,bachelor,engineer.His main research interests include big data and environmental informatization.
  • Supported by:
    National Natural Science Foundation of China(62271128), Sichuan Regional Innovation Cooperation Project(2020YFQ0018) and Sichuan Science and Technology Program(2022ZDZX0004,2023YFG0029,2023YFG0150,2023ZHCG004,2022YFG0212,2021YFS0391,2021YFG0027).

PDF (PC)

2518

摘要： 近年来,为了应对谣言广泛传播所带来的一系列社会问题,研究者开发了许多基于深度学习的谣言检测方法。虽然这些方法通过从传播结构中学习谣言的高级表征实现了较优的检测性能,但它们都忽略了在构造传播网络时边的不确定性,导致模型的可靠性降低,出现累积误差。针对该问题,提出了边推断增强对比学习的社交媒体谣言检测模型(Edge-Inference Con-trastive Learning,EICL)。首先,EICL基于消息转发(评论)时间戳为给定消息构建传播图;然后,利用新设计的边权重调整策略进行事件传播图数据增强以捕获传播结构边的不确定性;最后,利用对比学习方法解决原数据集本身存在的稀疏性问题,提高模型泛化能力。实验结果表明,与其他基准模型相比,模型EICL在公开数据集Twitter15和Twitter16上的准确率分别提高了2.0%和3.0%,证明其可显著提升社交媒体谣言检测效果。

关键词: 谣言检测, 对比学习, 数据增强, 因果推断

Abstract: In recent years,in order to deal with various social problems which are caused by the wide spreading of rumors,researchers have developed many deep learning-based rumor detection methods.Although these methods improve detection performance by learning the high-level representation of rumor from its propagation structure,they still suffer the problem of lower reliability and cumulative errors effect,due to the ignoring of edges’ uncertainty when constructing the propagation network.To address such a problem,this paper proposes the edge-inference contrastive learning(EIC) model.EICL first constructs a propagation graph based on timestamps of retweets(comments) for a given message.Then,it augments the event propagation graph to capture the edge uncertainty of the propagation structure by a newly designed edge-weight adjustment strategy.Finally,it employs the contrastive learning technique to solve the sparsity problem of the original dataset and improve the model generalization.Experimental results show that the accuracy of EICL is improved by 2.0% and 3.0% on Twitter15 and Twitter16,respectively,compared with other state-of-the-art baselines,which demonstrate that it can significantly improve the performance of rumor detection on social media.

Key words: Rumor detection, Contrastive learning, Data augmentation, Causal inference

中图分类号: 

  • TP181
[1]CASTILLO C,POBLETE B,MENDOZA M,et al.InformationCredibility on Twitter[C]//Proceedings of the 20th Interna-tional Conference on World Wide Web.2011:675-84.
[2]KWON S,CHA M,JUNG K,et al.Prominent Features of Ru-mor Propagation in Online Social Media[C]//2013 IEEE 13th International Conference on Data Mining.2013:1103-1108.
[3]VOSOUGHI S,ROY D,ARAL S.The Spread of True and False News Online[J].Science,2018,359(6380):1146-1151.
[4]MA J,GAO W,PRASENJIT M,et al.Detecting Rumors from Microblogs with Recurrent Neural Networks[C]//Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence.2016:3818-3824.
[5]KOTTETI C M M,DONG X,QIAN L.Ensemble deep learning on time-series representation of tweets for rumor detection in social media[J].Applied Sciences,2020,10(21):7541.
[6]KHOO L M S,CHIEU H L,QIAN Z,et al.Interpretable rumordetection in microblogs by attending to user interactions[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2020,34(5):8783-8790.
[7]BIAN T,XIAO X,XU T,et al.Rumor detection on social media with bi-directional graph convolutional networks[C]//Procee-dings of the AAAI Conference on Artificial Intelligence.2020,34(1):549-556.
[8]WEI P,XU N,MAO W.Modeling conversation structure andtemporal dynamics for jointly predicting rumor stance and vera-city[J].arXiv:1909.08211,2019.
[9]SHARMA K,QIAN F,JIANG H,et al.Combating fake news:A survey on identification and mitigation techniques[J].ACM Transactions on Intelligent Systems and Technology(TIST),2019,10(3):1-42.
[10]ZHOU X,ZAFARANI R.A survey of fake news:Fundamental theories,detection methods,and opportunities[J].ACM Computing Surveys(CSUR),2020,53(5):1-40.
[11]LIU X,NOURBAKHSH A,LI Q,et al.Real-time rumor debunking on twitter[C]//Proceedings of the 24th ACM International on Conference on Information and Knowledge Management.2015:1867-1870.
[12]SHELKE S,ATTAR V.Rumor Detection in Social NetworkBased on User,Content and Lexical Features[J].Multimedia Tools and Applications,2022,81(12):347-368.
[13]ZHANG H,QIAN S,FANG Q,et al.Multi-modal Meta Multi-task Learning for Social Media Rumor Detection[J].IEEE Transactions on Multimedia,2021,24,1449-1459.
[14]MA J,GAO W,WONG K.Rumor Detection on Twitter withTree-structured Recursive Neural Networks[C]//Proceedings of the 56th Annual Meeting of the Association for Computa-tional Linguistics.Association for Computa-tional Linguistics,2018:1980-1989.
[15]CHEN T,LI X,YIN H,et al.Call attention to rumors:Deep attention based recurrent neural networks for early rumor detection[C]//Pacific-Asia Conference on Knowledge Discovery and Data Mining.Cham:Springer,2018:40-52.
[16]MISHRA R,SETTY V.Sadhan:Hierarchical attention net-works to learn latent aspect embeddings for fake news detection[C]//Proceedings of the 2019 ACM SIGIR International Conference on Theory of Information Retrieval.2019:197-204.
[17]RAN H,JIA C,ZHANG P,et al.MGAT-ESM:Multi-channelgraph attention neural network with event-sharing module for rumor detection[J].Information Sciences,2022,592:402-416.
[18]NGUYEN V H,SUGIYAMA K,NAKOV P,et al.Fang:Leve-raging social context for fake news detection using graph representation[C]//Proceedings of the 29th ACM International Conference on Information & Knowledge Management.2020:1165-1174.
[19]WANG T,ISOLA P.Understanding contrastive representationlearning through alignment and uniformity on the hypersphere[C]//International Conference on Machine Learning.PMLR,2020:9929-9939.
[20]LIN H,MA J,CHENG M,et al.Rumor detection on twitter with claim-guided hierarchical graph attention networks[J].arXiv:2110.04522,2021.
[21]LIN H,YAN Y,CHEN G.Boosting Low-resource Intent Detection with In-scope Prototypical Networks[C]//ICASSP 2021－2021 IEEE International Conference on Acoustics.Speech and Signal Processing,2021:7623-7627.
[22]GAO T,YAO X,CHEN D.Simcse:Simple contrastive learning of sentence embeddings[J].arXiv:2104.08821,2021.
[23]MOHTARAMI M,GLASS J,NAKOV P.Contrastive language adaptation for cross-lingual stance detection[J].arXiv:1910.02076,2019.
[24]ZHUANG C,MA Q.Dual graph convolutional networks forgraph-based semi-supervised classification[C]//Proceedings of the 2018 World Wide Web Conference.2018:499-508.
[25]MA J,GAO W,WONG K.Detect Rumors in Microblog PostsUsing Propagation Structure Via Kernel Learning[C]//Procee-dings of the 55th Annual Meeting of the Association for Computational Linguistics.Association for Computational Linguistics,2017:708-717.
[26]WEI L,HU D,ZHOU W,et al.Towards propagation uncertainty:Edge-enhanced bayesian graph convolutional networks for rumor detection[J].arXiv:2107.11934,2021.
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