隨著社交網絡的發展，用戶能夠快速地獲取大量信息並搜尋感興趣的特定主題。然而，在這些龐大的信息量中，並非所有的信息都是準確的。因此，準確地檢測信息的真實性至關重要。以往的研究表明，即使是對傳播謠言的圖進行輕微的干擾，也能顯著改變預測結果。例如，簡單的操作如刪除一個節點或邊，可能會顛覆預測結果，將原本預測為假的圖變為真。因此，我們的目標是開發一種對此類故意干擾具有強大抵抗力的謠言預測模型。目前大多數的謠言檢測方法依賴於監督學習，需要大量的標註數據。與之相對，我們旨在無需謠言的明確標籤，通過無監督的方式實現可靠的預測。
我們在本文中提出了RUGRD，即Robust Unsupervised Graph Rumor Detection。我們採用了圖自動編碼器（Graph AutoEncoder）作為基本模型，並使用RGCN 作為編碼器來捕捉傳播的雙向性。GAE 的鏈接重建損失總和被用作我們判斷的依據。此外，我們還在模型中添加了特徵重建損失和群體聚類損失，以考慮謠言的語義和傳播模式。此外，我們將節點級對比學習和圖級對比學習整合到模型中，其目的是確保模型在面對輕微干擾時仍能保持穩健性和一致性，從而在不同條件下保持一致的判斷。最後，我們加入了基於課程的對比學習，讓模型能夠有效利用其自身的預測結果。
我們在三個不同的數據集上進行了實驗，結果顯示，我們的模型不僅在預測準確度上表現優異，且即使在圖中包含大量噪聲的情況下，也能保持穩健的性能。

With the advancement of social networks, users can access vast amounts of information quickly and search for specific topics of interest. However, not all information within this vast volume is accurate. Therefore, accurately detecting the veracity of information is crucial. Previous research has shown that even minor disturbances to a graph representing the spread of rumors can significantly alter prediction outcomes. For instance, simple actions like deleting a node or an edge can drastically change predictions, potentially flipping a graph originally predicted as False to be predicted as True. Hence, our goal is to develop a robust rumor prediction model resilient to such deliberate disturbances. Most current rumor detection methods rely on supervised learning, necessitating substantial annotated data. In contrast, we aim to achieve reliable predictions in an unsupervised manner without explicit labels for rumors.
We proposed RUGRD, which is a Robust Unsupervised Graph Rumor Detection, in this paper. We utilized Graph AutoEncoder as our basic model, and use RGCN as our encoder in order to capture two directional of the propagation. The total of Link reconstruct loss of GAE are used as the basis for our judgment. Besides, we we also add feature reconstruct loss and group clustering loss in our model in order to consider the semantic and spreading patterns of rumor. In addition, we integrate node-level contrastive learning and graph-level contrastive learning into our model. The objective is to ensure that our model remains robust and unaffected by minor disturbances, thereby maintaining consistent judgment across varying conditions. Finally, we incorporate curriculum-based contrastive learning to enable our model to utilize its own prediction results effectively.
We conducted experiments on three different datasets, and the results demonstrate that our model not only achieves superior prediction accuracy but also maintains robust performance even when the graph contains substantial noise.

[1] Tian Bian, Xi Xiao, Tingyang Xu, Peilin Zhao, Wenbing Huang, Yu Rong, and Junzhou Huang. Rumor detection on social media with bi-directional graph convolutional networks, 2020.
[2] Tianqi Chen and Carlos Guestrin. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ＇16. ACM, August 2016.
[3] Jingcan Duan, Siwei Wang, Pei Zhang, En Zhu, Jingtao Hu, Hu Jin, Yue Liu, and Zhibin Dong. Graph anomaly detection via multi-scale contrastive learning networks with augmented view, 2022.
[4] Lanting Fang, Kaiyu Feng, Kaiqi Zhao, Aiqun Hu, and Tao Li. Unsupervised rumor detection based on propagation tree vae. IEEE Transactions on Knowledge and Data Engineering, 35(10):10309–10323, 2023.
[5] Adam Goodge, Bryan Hooi, See Kiong Ng, and Wee Siong Ng. Lunar: Unifying local outlier detection methods via graph neural networks, 2021.
[6] R. Hadsell, S. Chopra, and Y. LeCun. Dimensionality reduction by learning an invariant mapping. In 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’06), volume 2, pages 1735–1742, 2006.
[7] Junwei He, Qianqian Xu, Yangbangyan Jiang, Zitai Wang, and Qingming Huang. Adagad: Anomaly-denoised autoencoders for graph anomaly detection, 2023.
[8] Kaiming He, Haoqi Fan, Yuxin Wu, Saining Xie, and Ross Girshick. Momentum contrast for unsupervised visual representation learning, 2020.
[9] M.A. Hearst, S.T. Dumais, E. Osuna, J. Platt, and B. Scholkopf. Support vector machines. IEEE Intelligent Systems and their Applications, 13(4):18–28, 1998.
[10] Geoffrey E Hinton and Richard Zemel. Autoencoders, minimum description length and helmholtz free energy. In J. Cowan, G. Tesauro, and J. Alspector, editors, Advances in Neural Information Processing Systems, volume 6. Morgan-Kaufmann, 1993.
[11] R Devon Hjelm, Alex Fedorov, Samuel Lavoie-Marchildon, Karan Grewal, Phil Bachman, Adam Trischler, and Yoshua Bengio. Learning deep representations by mutual information estimation and maximization, 2019.
[12] Zhenyu Hou, Yufei He, Yukuo Cen, Xiao Liu, Yuxiao Dong, Evgeny Kharlamov, and Jie Tang. Graphmae2: A decoding-enhanced masked self-supervised graph learner, 2023.
[13] Zhenyu Hou, Xiao Liu, Yukuo Cen, Yuxiao Dong, Hongxia Yang, Chunjie Wang, and Jie Tang. Graphmae: Self-supervised masked graph autoencoders, 2022.
[14] Diederik P Kingma and Max Welling. Auto-encoding variational bayes, 2022.
[15] Thomas N. Kipf and Max Welling. Variational graph auto-encoders, 2016.
[16] Thomas N. Kipf and Max Welling. Semi-supervised classification with graph convolutional networks, 2017.
[17] Dun Li, Haimei Guo, Zhenfei Wang, and Zhiyun Zheng. Unsupervised fake news detection based on autoencoder. IEEE Access, 9:29356–29365, 2021.
[18] Jindong Li, Qianli Xing, Qi Wang, and Yi Chang. CVTGAD: Simplified Transformer with Cross-View Attention for Unsupervised Graph-Level Anomaly Detection, page 185–200. Springer Nature Switzerland, 2023.
[19] Jintang Li, Ruofan Wu, Wangbin Sun, Liang Chen, Sheng Tian, Liang Zhu, Changhua Meng, Zibin Zheng, and Weiqiang Wang. What’s behind the mask: Understanding masked graph modeling for graph autoencoders, 2023.
[20] Shaohua Li, Weimin Li, Alex Munyole Luvembe, and Weiqin Tong. Graph contrastive learning with feature augmentation for rumor detection. IEEE Transactions on Computational Social Systems, pages 1–10, 2023.
[21] Hongzhan Lin, Jing Ma, Mingfei Cheng, Zhiwei Yang, Liangliang Chen, and Guang Chen. Rumor detection on Twitter with claim-guided hierarchical graph attention networks. In Marie-Francine Moens, Xuanjing Huang, Lucia Specia, and Scott Wen-tau Yih, editors, Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 10035–10047, Online and Punta Cana, Dominican Republic, November 2021. Association for Computational Linguistics.
[22] Fei Tony Liu, Kai Ming Ting, and Zhi-Hua Zhou. Isolation forest. In 2008 Eighth IEEE International Conference on Data Mining, pages 413–422, 2008.
[23] Yixin Liu, Kaize Ding, Huan Liu, and Shirui Pan. Good-d: On unsupervised graph out-of-distribution detection. In Proceedings of the Sixteenth ACM International Conference on Web Search and Data Mining, WSDM ’23, page 339–347, New York, NY, USA, 2023. Association for Computing Machinery.
[24] Guanghui Ma, Chunming Hu, Ling Ge, Junfan Chen, Hong Zhang, and Richong Zhang. Towards robust false information detection on social networks with contrastive learning. In Proceedings of the 31st ACM International Conference on Information & Knowledge Management, CIKM ’22, page 1441–1450, New York, NY, USA, 2022. Association for Computing Machinery.
[25] Jiachen Ma, Yong Liu, Meng Han, Chunqiang Hu, and Zhaojie Ju. Propagation structure fusion for rumor detection based on node-level contrastive learning. IEEE Transactions on Neural Networks and Learning Systems, pages 1–12, 2023.
[26] Jing Ma, Wei Gao, and Kam-Fai Wong. Detect rumors in microblog posts using propagation structure via kernel learning. In Regina Barzilay and Min-Yen Kan, editors, Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 708–717, Vancouver, Canada, July 2017. Association for Computational Linguistics.
[27] Jing Ma, Wei Gao, and Kam-Fai Wong. Rumor detection on Twitter with tree-structured recursive neural networks. In Iryna Gurevych and Yusuke Miyao, editors, Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1980–1989, Melbourne, Australia, July 2018. Association for Computational Linguistics.
[28] Jing Ma, Wei Gao, and Kam-Fai Wong. Detect rumors on twitter by promoting information campaigns with generative adversarial learning. In The World Wide Web Conference, WWW ’19, page 3049–3055, New York, NY, USA, 2019. Association for Computing Machinery.
[29] Joshua Robinson, Ching-Yao Chuang, Suvrit Sra, and Stefanie Jegelka. Contrastive learning with hard negative samples, 2021.
[30] David E. Rumelhart, Geoffrey E. Hinton, and Ronald J. Williams. Learning representations by back-propagating errors, 1986.
[31] Michael Schlichtkrull, Thomas N. Kipf, Peter Bloem, Rianne van den Berg, Ivan Titov, and Max Welling. Modeling relational data with graph convolutional networks, 2017.
[32] Yucheng Shi, Yushun Dong, Qiaoyu Tan, Jundong Li, and Ninghao Liu. Gigamae: Generalizable graph masked autoencoder via collaborative latent space reconstruction, 2023.
[33] Fan-Yun Sun, Jordan Hoffmann, Vikas Verma, and Jian Tang. Infograph: Unsupervised and semi-supervised graph-level representation learning via mutual information maximization, 2020.
[34] Tiening Sun, Zhong Qian, Sujun Dong, Peifeng Li, and Qiaoming Zhu. Rumor detection on social media with graph adversarial contrastive learning. In Proceedings of the ACM Web Conference 2022, WWW ’22, page 2789–2797, New York, NY, USA, 2022. Association for Computing Machinery.
[35] Qiaoyu Tan, Ninghao Liu, Xiao Huang, Soo-Hyun Choi, Li Li, Rui Chen, and Xia Hu. S2gae: Self-supervised graph autoencoders are generalizable learners with graph masking. In Proceedings of the Sixteenth ACM International Conference on Web Search and Data Mining, WSDM ’23, page 787–795, New York, NY, USA, 2023. Association for Computing Machinery.
[36] Xiang Tao, Liang Wang, Qiang Liu, Shu Wu, and Liang Wang. Semantic evolvement enhanced graph autoencoder for rumor detection, 2024.
[37] Xiang Tao, Mingqing Zhang, Qiang Liu, Shu Wu, and Liang Wang. Out-of-distribution rumor detection via test-time adaptation, 2024.
[38] Petar Veličković, William Fedus, William L. Hamilton, Pietro Liò, Yoshua Bengio, and R Devon Hjelm. Deep graph infomax, 2018.
[39] Herun Wan, Shangbin Feng, Zhaoxuan Tan, Heng Wang, Yulia Tsvetkov, and Minnan Luo. Dell: Generating reactions and explanations for llm-based misinformation detection, 2024.
[40] Yanling Wang, Jing Zhang, Shasha Guo, Hongzhi Yin, Cuiping Li, and Hong Chen. Decoupling representation learning and classification for gnn-based anomaly detection. Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, 2021.
[41] Minghao Xu, Hang Wang, Bingbing Ni, Hongyu Guo, and Jian Tang. Self-supervised graph-level representation learning with local and global structure, 2021.
[42] Xiaoyu Yang, Yuefei Lyu, Tian Tian, Yifei Liu, Yudong Liu, and Xi Zhang. Rumor detection on social media with graph structured adversarial learning. In Christian Bessiere, editor, Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence, IJCAI-20, pages 1417–1423. International Joint Conferences on Artificial Intelligence Organization, 7 2020. Main track.
[43] Yaowen Ye, Lianghao Xia, and Chao Huang. Graph masked autoencoder for sequential recommendation. In Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR ＇23. ACM, July 2023.
[44] Jing Yi and Zhenzhong Chen. Multi-modal variational graph auto-encoder for recommendation systems. IEEE Transactions on Multimedia, 24:1067–1079, 2022.
[45] Shu Yin, Chao Gao, and Zhen Wang. Gamc: An unsupervised method for fake news detection using graph autoencoder with masking, 2023.
[46] Yuning You, Tianlong Chen, Yongduo Sui, Ting Chen, Zhangyang Wang, and Yang Shen. Graph contrastive learning with augmentations, 2021.
[47] Feng Yu, Qiang Liu, Shu Wu, Liang Wang, and Tieniu Tan. A convolutional approach for misinformation identification. In Proceedings of the 26th International Joint Conference on Artificial Intelligence, IJCAI’17, page 3901–3907. AAAI Press, 2017.
[48] Chunyuan Yuan, Qianwen Ma, Wei Zhou, Jizhong Han, and Songlin Hu. Jointly embedding the local and global relations of heterogeneous graph for rumor detection, 2019.
[49] Huaiwen Zhang, Xinxin Liu, Qing Yang, Yang Yang, Fan Qi, Shengsheng Qian, and Changsheng Xu. T3rd: Test-time training for rumor detection on social media. In Proceedings of the ACM on Web Conference 2024, WWW ’24, page 2407–2416, New York, NY, USA, 2024. Association for Computing Machinery.
[50] Yue Zhao, Xiyang Hu, Cheng Cheng, Cong Wang, Changlin Wan, Wen Wang, Jianing Yang, Haoping Bai, Zheng Li, Cao Xiao, Yunlong Wang, Zhi Qiao, Jimeng Sun, and Leman Akoglu. Suod: Accelerating large-scale unsupervised heterogeneous outlier detection, 2021.