Publications-Theses
Article View/Open
Publication Export
-
題名 以深度動態卷積神經網路實施多重任務學習偵測假新聞
Deep Dynamic Convolutional Neural Network with Multi-Task Learning for Fake News Detection作者 林佑駿
Lin, Yu-Chun貢獻者 胡毓忠
Hu, Yuh-Jong
林佑駿
Lin, Yu-Chun關鍵詞 假新聞
深度學習
社群媒體
動態卷積神經網路
多重任務學習
Fake News
Detection
Deep Learning
Social Media
Dynamic CNN
Multi-Task Learning日期 2020 上傳時間 2-Mar-2020 11:38:27 (UTC+8) 摘要 傳統的假新聞偵測主要區分為知識庫比對、專家人工辨識與特徵機器學習等3大方式，但是隨著資料數據的日益龐大、新聞來源的多樣化以及惡意變造新聞的手法推層出新，傳統假新聞偵測方法已出現瓶頸，逐漸不敷現況使用，為了突破此一困境，於是出現以深度學習找尋未知特徵的偵測方式。\n以往深度學習受限於硬體效能，不易針對模型的調整與優化進行全方位驗測，所幸隨著科技進步與莫爾定律推演，硬體效能持續以指數性程度成長，進而使深度學習的研究邁向了全新領域。\n本論文除了研究如何以深度學習中的深度動態卷積神經網路進行假新聞偵\n測外，同時也探討超參數在深度學習中對於優化的影響及資料集特徵在模型中所扮演的角色。運用多重任務學習框架，將推文立場、假新聞偵測與假新聞驗證等3個任務相互搭配，分析各任務彼此之間的關連影響。另針對深度動態卷積神經網路在處理假新聞偵測應用問題上的特性進行分析。
Traditional fake news detection is mainly divided into three major methods, knowledge based comparison, expert manual identification, and feature machine learning. With the increasing data, the diversification of news sources, and the malicious method of altering news, the traditional methods of detecting fake news has become a bottleneck, and it is gradually inadequate to use it. To break through this dilemma, there is a detection method that uses deep learning to find unknown features.\nIn the past, deep learning was limited by hardware performance, and it was not easy to conduct comprehensive testing for model adjustment and optimization.\nFortunately, with the advancement of science and technology and Moore`s Law, hardware performance continued to grow exponentially, deep learning towards a whole new field.\nIn addition to studying how to detect fake news with deep dynamic convolutional neural networks in deep learning, this paper also explores the impact of\nhyperparameters on deep learning optimization and the role of data set features in the model. Besides, a multi-task learning framework is used to match three tasks, such as tweet position, fake news detection, and fake news verification, to analyze the impact of each task on each other. It also analyzes the characteristics of the deep dynamic convolutional neural network in dealing with the application of fake news detection.參考文獻 [1] N. Kalchbrenner, E. Grefenstette, and P. Blunsom, "A convolutional neural\nnetwork for modelling sentences," arXiv preprint arXiv:1404.2188, 2014.\n[2] Z. Tufekci, "It’s the (democracy-poisoning) Golden Age of free speech,"\nWIRED. Accessed May, vol. 20, p. 2018, 2018.\n[3] E. Hunt, "What is fake news? How to spot it and what you can do to stop it,"\nThe Guardian, vol. 17, 2016.\n[4] K. Shu, A. Sliva, S. Wang, J. Tang, and H. Liu, "Fake news detection on\nsocial media: A data mining perspective," ACM SIGKDD Explorations\nNewsletter, vol. 19, no. 1, pp. 22-36, 2017.\n[5] V. Rubin, N. Conroy, Y. Chen, and S. Cornwell, "Fake news or truth? using\nsatirical cues to detect potentially misleading news," in Proceedings of the\nsecond workshop on computational approaches to deception detection, 2016,\npp. 7-17.\n[6] M. Balmas, "When fake news becomes real: Combined exposure to multiple\nnews sources and political attitudes of inefficacy, alienation, and cynicism,"\nCommunication Research, vol. 41, no. 3, pp. 430-454, 2014.\n[7] E. Ferrara, O. Varol, C. Davis, F. Menczer, and A. Flammini, "The rise of\nsocial bots," Communications of the ACM, vol. 59, no. 7, pp. 96-104, 2016.\n[8] J. Cheng, M. Bernstein, C. Danescu-Niculescu-Mizil, and J. Leskovec,\n"Anyone can become a troll: Causes of trolling behavior in online\ndiscussions," in Proceedings of the 2017 ACM conference on computer\nsupported cooperative work and social computing, 2017: ACM, pp. 1217-\n1230.\n[9] Z. Chu, S. Gianvecchio, H. Wang, and S. Jajodia, "Detecting automation of\ntwitter accounts: Are you a human, bot, or cyborg?," IEEE Transactions on\nDependable and Secure Computing, vol. 9, no. 6, pp. 811-824, 2012.\n[10] M. Del Vicario et al., "Echo chambers: Emotional contagion and group\npolarization on facebook," Scientific reports, vol. 6, p. 37825, 2016.\n[11] W. Quattrociocchi, A. Scala, and C. R. Sunstein, "Echo chambers on\nFacebook," Available at SSRN 2795110, 2016.\n[12] C. Paul and M. Matthews, "The Russian “firehose of falsehood” propaganda\nmodel," Rand Corporation, pp. 2-7, 2016.\n[13] Y. Chen, N. J. Conroy, and V. L. Rubin, "Misleading online content:\nRecognizing clickbait as false news," in Proceedings of the 2015 ACM on\nWorkshop on Multimodal Deception Detection, 2015: ACM, pp. 15-19.\n[14] J. Fürnkranz, "A study using n-gram features for text categorization," Austrian\nResearch Institute for Artifical Intelligence, vol. 3, no. 1998, pp. 1-10, 1998.\n[15] M. Potthast, J. Kiesel, K. Reinartz, J. Bevendorff, and B. Stein, "A stylometric\ninquiry into hyperpartisan and fake news," arXiv preprint arXiv:1702.05638,\n2017.\n[16] S. Afroz, M. Brennan, and R. Greenstadt, "Detecting hoaxes, frauds, and\ndeception in writing style online," in 2012 IEEE Symposium on Security and\nPrivacy, 2012: IEEE, pp. 461-475.\n[17] C. Castillo, M. Mendoza, and B. Poblete, "Information credibility on twitter,"\nin Proceedings of the 20th international conference on World wide web, 2011:\nACM, pp. 675-684.\n[18] F. Yang, Y. Liu, X. Yu, and M. Yang, "Automatic detection of rumor on\nSina Weibo," in Proceedings of the ACM SIGKDD Workshop on Mining Data\nSemantics, 2012: ACM, p. 13.\n[19] J. Ma, W. Gao, Z. Wei, Y. Lu, and K.-F. Wong, "Detect rumors using time\nseries of social context information on microblogging websites," in\nProceedings of the 24th ACM International on Conference on Information and\nKnowledge Management, 2015: ACM, pp. 1751-1754.\n[20] S. Kwon, M. Cha, K. Jung, W. Chen, and Y. Wang, "Prominent features of\nrumor propagation in online social media," in 2013 IEEE 13th International\nConference on Data Mining, 2013: IEEE, pp. 1103-1108.\n[21] N. Ruchansky, S. Seo, and Y. Liu, "Csi: A hybrid deep model for fake news\ndetection," in Proceedings of the 2017 ACM on Conference on Information\nand Knowledge Management, 2017: ACM, pp. 797-806.\n[22] Z. Jin, J. Cao, Y. Zhang, and J. Luo, "News verification by exploiting\nconflicting social viewpoints in microblogs," in Thirtieth AAAI Conference on\nArtificial Intelligence, 2016.\n[23] J. Ma et al., "Detecting rumors from microblogs with recurrent neural\nnetworks," in Ijcai, 2016, pp. 3818-3824.\n[24] E. Tacchini, G. Ballarin, M. L. Della Vedova, S. Moret, and L. de Alfaro,\n"Some like it hoax: Automated fake news detection in social networks," arXiv\npreprint arXiv:1704.07506, 2017.\n[25] A. Zubiaga, A. Aker, K. Bontcheva, M. Liakata, and R. Procter, "Detection\nand resolution of rumours in social media: A survey," ACM Computing\nSurveys (CSUR), vol. 51, no. 2, p. 32, 2018.\n[26] A. Vlachos and S. Riedel, "Fact checking: Task definition and dataset\nconstruction," in Proceedings of the ACL 2014 Workshop on Language\nTechnologies and Computational Social Science, 2014, pp. 18-22.\n[27] N. Hassan, C. Li, and M. Tremayne, "Detecting check-worthy factual claims\nin presidential debates," in Proceedings of the 24th acm international on\nconference on information and knowledge management, 2015: ACM, pp.\n1835-1838.\n[28] M. Banko, M. J. Cafarella, S. Soderland, M. Broadhead, and O. Etzioni,\n"Open information extraction from the web," in Ijcai, 2007, vol. 7, pp. 2670-\n2676.\n[29] A. Magdy and N. Wanas, "Web-based statistical fact checking of textual\ndocuments," in Proceedings of the 2nd international workshop on Search and\nmining user-generated contents, 2010: ACM, pp. 103-110.\n[30] Y. Wu, P. K. Agarwal, C. Li, J. Yang, and C. Yu, "Toward computational\nfact-checking," Proceedings of the VLDB Endowment, vol. 7, no. 7, pp. 589-\n600, 2014.\n[31] G. L. Ciampaglia, P. Shiralkar, L. M. Rocha, J. Bollen, F. Menczer, and A.\nFlammini, "Computational fact checking from knowledge networks," PloS\none, vol. 10, no. 6, p. e0128193, 2015.\n[32] B. Shi and T. Weninger, "Fact checking in heterogeneous information\nnetworks," in Proceedings of the 25th International Conference Companion\non World Wide Web, 2016: International World Wide Web Conferences\nSteering Committee, pp. 101-102.\n[33] Z. Zhao, P. Resnick, and Q. Mei, "Enquiring minds: Early detection of\nrumors in social media from enquiry posts," in Proceedings of the 24th\nInternational Conference on World Wide Web, 2015: International World\nWide Web Conferences Steering Committee, pp. 1395-1405.\n[34] A. Zubiaga, M. Liakata, and R. Procter, "Exploiting context for rumour\ndetection in social media," in International Conference on Social Informatics,\n2017: Springer, pp. 109-123.\n[35] S. Kwon, M. Cha, and K. Jung, "Rumor detection over varying time\nwindows," PloS one, vol. 12, no. 1, p. e0168344, 2017.\n[36] T. Chen, X. Li, H. Yin, and J. Zhang, "Call attention to rumors: Deep\nattention based recurrent neural networks for early rumor detection," in Pacific-Asia Conference on Knowledge Discovery and Data Mining, 2018:\nSpringer, pp. 40-52.\n[37] M. Mendoza, B. Poblete, and C. Castillo, "Twitter under crisis: Can we trust\nwhat we RT?," in Proceedings of the first workshop on social media analytics,\n2010: ACM, pp. 71-79.\n[38] R. Procter, F. Vis, and A. Voss, "Reading the riots on Twitter: methodological\ninnovation for the analysis of big data," International journal of social\nresearch methodology, vol. 16, no. 3, pp. 197-214, 2013.\n[39] L. Derczynski et al., "PHEME: Computing Veracity—the Fourth Challenge\nof Big Social Data," in Proceedings of the Extended Semantic Web\nConference EU Project Networking session (ESCW-PN), 2015.\n[40] M. Lukasik, P. Srijith, D. Vu, K. Bontcheva, A. Zubiaga, and T. Cohn,\n"Hawkes processes for continuous time sequence classification: an application\nto rumour stance classification in twitter," in Proceedings of the 54th Annual\nMeeting of the Association for Computational Linguistics (Volume 2: Short\nPapers), 2016, vol. 2, pp. 393-398.\n[41] A. Zubiaga et al., "Discourse-aware rumour stance classification in social\nmedia using sequential classifiers," Information Processing & Management,\nvol. 54, no. 2, pp. 273-290, 2018.\n[42] W. Y. Wang, "" liar, liar pants on fire": A new benchmark dataset for fake\nnews detection," arXiv preprint arXiv:1705.00648, 2017.\n[43] G. Giasemidis et al., "Determining the veracity of rumours on Twitter," in\nInternational Conference on Social Informatics, 2016: Springer, pp. 185-205.\n[44] C. Boididou, S. Papadopoulos, Y. Kompatsiaris, S. Schifferes, and N.\nNewman, "Challenges of computational verification in social multimedia," in\nProceedings of the 23rd International Conference on World Wide Web, 2014:\nACM, pp. 743-748.\n[45] C. Boididou, S. Papadopoulos, L. Apostolidis, and Y. Kompatsiaris,\n"Learning to detect misleading content on twitter," in Proceedings of the 2017\nACM on International Conference on Multimedia Retrieval, 2017: ACM, pp.\n278-286.\n[46] O. Russakovsky et al., "Imagenet large scale visual recognition challenge,"\nInternational journal of computer vision, vol. 115, no. 3, pp. 211-252, 2015.\n[47] Y. Kim, "Convolutional neural networks for sentence classification," arXiv\npreprint arXiv:1408.5882, 2014.\n[48] I. Solaiman et al., "Release strategies and the social impacts of language\nmodels," arXiv preprint arXiv:1908.09203, 2019\n[49] J. Baxter, "A Bayesian/information theoretic model of learning to learn via\nmultiple task sampling," Machine learning, vol. 28, no. 1, pp. 7-39, 1997.\n[50] R. Collobert and J. Weston, "A unified architecture for natural language\nprocessing: Deep neural networks with multitask learning," in Proceedings of\nthe 25th international conference on Machine learning, 2008: ACM, pp. 160-\n167.\n[51] L. Derczynski, K. Bontcheva, M. Liakata, R. Procter, G. W. S. Hoi, and A.\nZubiaga, "SemEval-2017 Task 8: RumourEval: Determining rumour veracity\nand support for rumours," arXiv preprint arXiv:1704.05972, 2017.\n[52] R. Caruana, "Multitask learning: A knowledge-based source of inductive bias.\nMachine Learning," 1997.\n[53] S. Ruder, "An overview of multi-task learning in deep neural networks," arXiv\npreprint arXiv:1706.05098, 2017.\n[54] E. Kochkina, M. Liakata, and A. Zubiaga, "All-in-one: Multi-task learning for\nrumour verification," arXiv preprint arXiv:1806.03713, 2018.\n[55] O. Enayet and S. R. El-Beltagy, "NileTMRG at SemEval-2017 Task 8:\nDetermining Rumour and Veracity Support for Rumours on Twitter," in\nProceedings of the 11th International Workshop on Semantic Evaluation\n(SemEval-2017), 2017, pp. 470-474.\n[56] A. Zubiaga, M. Liakata, R. Procter, G. W. S. Hoi, and P. Tolmie, "Analysing\nhow people orient to and spread rumours in social media by looking at\nconversational threads," PloS one, vol. 11, no. 3, p. e0150989, 2016.\n[57] E. Kochkina, M. Liakata, and I. Augenstein, "Turing at semeval-2017 task 8:\nSequential approach to rumour stance classification with branch-lstm," arXiv\npreprint arXiv:1704.07221, 2017.\n[58] T. Mikolov, K. Chen, G. Corrado, and J. Dean, "Efficient estimation of word\nrepresentations in vector space," arXiv preprint arXiv:1301.3781, 2013.\n[59] D. Britz, "Understanding convolutional neural networks for NLP," URL:\nhttp://www. wildml. com/2015/11/understanding-convolutionalneuralnetworks-for-nlp/(visited on 11/07/2015), 2015.\n[60] T. Perry, "Convolutional methods for text," ed: Medium, 2017. 描述 碩士
國立政治大學
資訊科學系碩士在職專班
106971004資料來源 http://thesis.lib.nccu.edu.tw/record/#G0106971004 資料類型 thesis dc.contributor.advisor 胡毓忠 zh_TW dc.contributor.advisor Hu, Yuh-Jong en_US dc.contributor.author (Authors) 林佑駿 zh_TW dc.contributor.author (Authors) Lin, Yu-Chun en_US dc.creator (作者) 林佑駿 zh_TW dc.creator (作者) Lin, Yu-Chun en_US dc.date (日期) 2020 en_US dc.date.accessioned 2-Mar-2020 11:38:27 (UTC+8) - dc.date.available 2-Mar-2020 11:38:27 (UTC+8) - dc.date.issued (上傳時間) 2-Mar-2020 11:38:27 (UTC+8) - dc.identifier (Other Identifiers) G0106971004 en_US dc.identifier.uri (URI) https://ah.lib.nccu.edu.tw/item?item_id=147816 - dc.description (描述) 碩士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 資訊科學系碩士在職專班 zh_TW dc.description (描述) 106971004 zh_TW dc.description.abstract (摘要) 傳統的假新聞偵測主要區分為知識庫比對、專家人工辨識與特徵機器學習等3大方式，但是隨著資料數據的日益龐大、新聞來源的多樣化以及惡意變造新聞的手法推層出新，傳統假新聞偵測方法已出現瓶頸，逐漸不敷現況使用，為了突破此一困境，於是出現以深度學習找尋未知特徵的偵測方式。\n以往深度學習受限於硬體效能，不易針對模型的調整與優化進行全方位驗測，所幸隨著科技進步與莫爾定律推演，硬體效能持續以指數性程度成長，進而使深度學習的研究邁向了全新領域。\n本論文除了研究如何以深度學習中的深度動態卷積神經網路進行假新聞偵\n測外，同時也探討超參數在深度學習中對於優化的影響及資料集特徵在模型中所扮演的角色。運用多重任務學習框架，將推文立場、假新聞偵測與假新聞驗證等3個任務相互搭配，分析各任務彼此之間的關連影響。另針對深度動態卷積神經網路在處理假新聞偵測應用問題上的特性進行分析。 zh_TW dc.description.abstract (摘要) Traditional fake news detection is mainly divided into three major methods, knowledge based comparison, expert manual identification, and feature machine learning. With the increasing data, the diversification of news sources, and the malicious method of altering news, the traditional methods of detecting fake news has become a bottleneck, and it is gradually inadequate to use it. To break through this dilemma, there is a detection method that uses deep learning to find unknown features.\nIn the past, deep learning was limited by hardware performance, and it was not easy to conduct comprehensive testing for model adjustment and optimization.\nFortunately, with the advancement of science and technology and Moore`s Law, hardware performance continued to grow exponentially, deep learning towards a whole new field.\nIn addition to studying how to detect fake news with deep dynamic convolutional neural networks in deep learning, this paper also explores the impact of\nhyperparameters on deep learning optimization and the role of data set features in the model. Besides, a multi-task learning framework is used to match three tasks, such as tweet position, fake news detection, and fake news verification, to analyze the impact of each task on each other. It also analyzes the characteristics of the deep dynamic convolutional neural network in dealing with the application of fake news detection. en_US dc.description.tableofcontents 目錄\n摘要 i\nAbstract ii\n致謝 iii\n目錄 iv\n表目錄 vii\n圖目錄 viii\n第一章 導論 1\n1.1 研究動機 1\n1.2 研究目的 3\n1.3 研究貢獻 4\n第二章 研究背景 6\n2.1 假新聞定義 6\n2.1.1 社群媒體中的假新聞 7\n2.2 假新聞特徵 8\n2.2.1 社群媒體假新聞特徵 10\n2.3 假新聞分類方法 12\n2.3.1 假新聞偵測 14\n2.3.2 推文立場 14\n2.3.3 假新聞驗證 15\n2.4 卷積神經網路 16\n2.4.1 卷積神經網路應用於文本分類 17\n2.4.2 動態卷積神經網路 18\n2.5 自然語言處理 20\n2.6 多重任務學習 22\n第三章 相關研究 27\n3.1 假新聞分類相關研究 27\n第四章 研究方法與架構 28\n4.1 假新聞資料集 29\n4.1.1 PHEME 資料集 30\n4.1.2 RumourEval 資料集 32\n4.1.3 資料前處理 34\n4.2 模型驗證 34\n4.3 動態卷積網路建模 36\n4.4 多重任務學習 36\n第五章 研究實作與結果 38\n5.1 實驗環境 38\n5.2 實做過程 39\n5.3 實驗結果 40\n5.3.1 多任務深度動態卷積網路學習 40\n5.3.2 各類模型比較 46\n5.4 分析研討 48\n5.4.1 資料集規模 48\n5.4.2 動態池化參數 49\n5.4.3 多重任務學習 49\n5.4.4 深度動態卷積神經網路模型特色 50\n第六章 結論與未來展望 52\n6.1 研究結論 52\n6.2 未來展望 53\n參考文獻 55 zh_TW dc.format.extent 1858156 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0106971004 en_US dc.subject (關鍵詞) 假新聞 zh_TW dc.subject (關鍵詞) 深度學習 zh_TW dc.subject (關鍵詞) 社群媒體 zh_TW dc.subject (關鍵詞) 動態卷積神經網路 zh_TW dc.subject (關鍵詞) 多重任務學習 zh_TW dc.subject (關鍵詞) Fake News en_US dc.subject (關鍵詞) Detection en_US dc.subject (關鍵詞) Deep Learning en_US dc.subject (關鍵詞) Social Media en_US dc.subject (關鍵詞) Dynamic CNN en_US dc.subject (關鍵詞) Multi-Task Learning en_US dc.title (題名) 以深度動態卷積神經網路實施多重任務學習偵測假新聞 zh_TW dc.title (題名) Deep Dynamic Convolutional Neural Network with Multi-Task Learning for Fake News Detection en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) [1] N. Kalchbrenner, E. Grefenstette, and P. Blunsom, "A convolutional neural\nnetwork for modelling sentences," arXiv preprint arXiv:1404.2188, 2014.\n[2] Z. Tufekci, "It’s the (democracy-poisoning) Golden Age of free speech,"\nWIRED. Accessed May, vol. 20, p. 2018, 2018.\n[3] E. Hunt, "What is fake news? How to spot it and what you can do to stop it,"\nThe Guardian, vol. 17, 2016.\n[4] K. Shu, A. Sliva, S. Wang, J. Tang, and H. Liu, "Fake news detection on\nsocial media: A data mining perspective," ACM SIGKDD Explorations\nNewsletter, vol. 19, no. 1, pp. 22-36, 2017.\n[5] V. Rubin, N. Conroy, Y. Chen, and S. Cornwell, "Fake news or truth? using\nsatirical cues to detect potentially misleading news," in Proceedings of the\nsecond workshop on computational approaches to deception detection, 2016,\npp. 7-17.\n[6] M. Balmas, "When fake news becomes real: Combined exposure to multiple\nnews sources and political attitudes of inefficacy, alienation, and cynicism,"\nCommunication Research, vol. 41, no. 3, pp. 430-454, 2014.\n[7] E. Ferrara, O. Varol, C. Davis, F. Menczer, and A. Flammini, "The rise of\nsocial bots," Communications of the ACM, vol. 59, no. 7, pp. 96-104, 2016.\n[8] J. Cheng, M. Bernstein, C. Danescu-Niculescu-Mizil, and J. Leskovec,\n"Anyone can become a troll: Causes of trolling behavior in online\ndiscussions," in Proceedings of the 2017 ACM conference on computer\nsupported cooperative work and social computing, 2017: ACM, pp. 1217-\n1230.\n[9] Z. Chu, S. Gianvecchio, H. Wang, and S. Jajodia, "Detecting automation of\ntwitter accounts: Are you a human, bot, or cyborg?," IEEE Transactions on\nDependable and Secure Computing, vol. 9, no. 6, pp. 811-824, 2012.\n[10] M. Del Vicario et al., "Echo chambers: Emotional contagion and group\npolarization on facebook," Scientific reports, vol. 6, p. 37825, 2016.\n[11] W. Quattrociocchi, A. Scala, and C. R. Sunstein, "Echo chambers on\nFacebook," Available at SSRN 2795110, 2016.\n[12] C. Paul and M. Matthews, "The Russian “firehose of falsehood” propaganda\nmodel," Rand Corporation, pp. 2-7, 2016.\n[13] Y. Chen, N. J. Conroy, and V. L. Rubin, "Misleading online content:\nRecognizing clickbait as false news," in Proceedings of the 2015 ACM on\nWorkshop on Multimodal Deception Detection, 2015: ACM, pp. 15-19.\n[14] J. Fürnkranz, "A study using n-gram features for text categorization," Austrian\nResearch Institute for Artifical Intelligence, vol. 3, no. 1998, pp. 1-10, 1998.\n[15] M. Potthast, J. Kiesel, K. Reinartz, J. Bevendorff, and B. Stein, "A stylometric\ninquiry into hyperpartisan and fake news," arXiv preprint arXiv:1702.05638,\n2017.\n[16] S. Afroz, M. Brennan, and R. Greenstadt, "Detecting hoaxes, frauds, and\ndeception in writing style online," in 2012 IEEE Symposium on Security and\nPrivacy, 2012: IEEE, pp. 461-475.\n[17] C. Castillo, M. Mendoza, and B. Poblete, "Information credibility on twitter,"\nin Proceedings of the 20th international conference on World wide web, 2011:\nACM, pp. 675-684.\n[18] F. Yang, Y. Liu, X. Yu, and M. Yang, "Automatic detection of rumor on\nSina Weibo," in Proceedings of the ACM SIGKDD Workshop on Mining Data\nSemantics, 2012: ACM, p. 13.\n[19] J. Ma, W. Gao, Z. Wei, Y. Lu, and K.-F. Wong, "Detect rumors using time\nseries of social context information on microblogging websites," in\nProceedings of the 24th ACM International on Conference on Information and\nKnowledge Management, 2015: ACM, pp. 1751-1754.\n[20] S. Kwon, M. Cha, K. Jung, W. Chen, and Y. Wang, "Prominent features of\nrumor propagation in online social media," in 2013 IEEE 13th International\nConference on Data Mining, 2013: IEEE, pp. 1103-1108.\n[21] N. Ruchansky, S. Seo, and Y. Liu, "Csi: A hybrid deep model for fake news\ndetection," in Proceedings of the 2017 ACM on Conference on Information\nand Knowledge Management, 2017: ACM, pp. 797-806.\n[22] Z. Jin, J. Cao, Y. Zhang, and J. Luo, "News verification by exploiting\nconflicting social viewpoints in microblogs," in Thirtieth AAAI Conference on\nArtificial Intelligence, 2016.\n[23] J. Ma et al., "Detecting rumors from microblogs with recurrent neural\nnetworks," in Ijcai, 2016, pp. 3818-3824.\n[24] E. Tacchini, G. Ballarin, M. L. Della Vedova, S. Moret, and L. de Alfaro,\n"Some like it hoax: Automated fake news detection in social networks," arXiv\npreprint arXiv:1704.07506, 2017.\n[25] A. Zubiaga, A. Aker, K. Bontcheva, M. Liakata, and R. Procter, "Detection\nand resolution of rumours in social media: A survey," ACM Computing\nSurveys (CSUR), vol. 51, no. 2, p. 32, 2018.\n[26] A. Vlachos and S. Riedel, "Fact checking: Task definition and dataset\nconstruction," in Proceedings of the ACL 2014 Workshop on Language\nTechnologies and Computational Social Science, 2014, pp. 18-22.\n[27] N. Hassan, C. Li, and M. Tremayne, "Detecting check-worthy factual claims\nin presidential debates," in Proceedings of the 24th acm international on\nconference on information and knowledge management, 2015: ACM, pp.\n1835-1838.\n[28] M. Banko, M. J. Cafarella, S. Soderland, M. Broadhead, and O. Etzioni,\n"Open information extraction from the web," in Ijcai, 2007, vol. 7, pp. 2670-\n2676.\n[29] A. Magdy and N. Wanas, "Web-based statistical fact checking of textual\ndocuments," in Proceedings of the 2nd international workshop on Search and\nmining user-generated contents, 2010: ACM, pp. 103-110.\n[30] Y. Wu, P. K. Agarwal, C. Li, J. Yang, and C. Yu, "Toward computational\nfact-checking," Proceedings of the VLDB Endowment, vol. 7, no. 7, pp. 589-\n600, 2014.\n[31] G. L. Ciampaglia, P. Shiralkar, L. M. Rocha, J. Bollen, F. Menczer, and A.\nFlammini, "Computational fact checking from knowledge networks," PloS\none, vol. 10, no. 6, p. e0128193, 2015.\n[32] B. Shi and T. Weninger, "Fact checking in heterogeneous information\nnetworks," in Proceedings of the 25th International Conference Companion\non World Wide Web, 2016: International World Wide Web Conferences\nSteering Committee, pp. 101-102.\n[33] Z. Zhao, P. Resnick, and Q. Mei, "Enquiring minds: Early detection of\nrumors in social media from enquiry posts," in Proceedings of the 24th\nInternational Conference on World Wide Web, 2015: International World\nWide Web Conferences Steering Committee, pp. 1395-1405.\n[34] A. Zubiaga, M. Liakata, and R. Procter, "Exploiting context for rumour\ndetection in social media," in International Conference on Social Informatics,\n2017: Springer, pp. 109-123.\n[35] S. Kwon, M. Cha, and K. Jung, "Rumor detection over varying time\nwindows," PloS one, vol. 12, no. 1, p. e0168344, 2017.\n[36] T. Chen, X. Li, H. Yin, and J. Zhang, "Call attention to rumors: Deep\nattention based recurrent neural networks for early rumor detection," in Pacific-Asia Conference on Knowledge Discovery and Data Mining, 2018:\nSpringer, pp. 40-52.\n[37] M. Mendoza, B. Poblete, and C. Castillo, "Twitter under crisis: Can we trust\nwhat we RT?," in Proceedings of the first workshop on social media analytics,\n2010: ACM, pp. 71-79.\n[38] R. Procter, F. Vis, and A. Voss, "Reading the riots on Twitter: methodological\ninnovation for the analysis of big data," International journal of social\nresearch methodology, vol. 16, no. 3, pp. 197-214, 2013.\n[39] L. Derczynski et al., "PHEME: Computing Veracity—the Fourth Challenge\nof Big Social Data," in Proceedings of the Extended Semantic Web\nConference EU Project Networking session (ESCW-PN), 2015.\n[40] M. Lukasik, P. Srijith, D. Vu, K. Bontcheva, A. Zubiaga, and T. Cohn,\n"Hawkes processes for continuous time sequence classification: an application\nto rumour stance classification in twitter," in Proceedings of the 54th Annual\nMeeting of the Association for Computational Linguistics (Volume 2: Short\nPapers), 2016, vol. 2, pp. 393-398.\n[41] A. Zubiaga et al., "Discourse-aware rumour stance classification in social\nmedia using sequential classifiers," Information Processing & Management,\nvol. 54, no. 2, pp. 273-290, 2018.\n[42] W. Y. Wang, "" liar, liar pants on fire": A new benchmark dataset for fake\nnews detection," arXiv preprint arXiv:1705.00648, 2017.\n[43] G. Giasemidis et al., "Determining the veracity of rumours on Twitter," in\nInternational Conference on Social Informatics, 2016: Springer, pp. 185-205.\n[44] C. Boididou, S. Papadopoulos, Y. Kompatsiaris, S. Schifferes, and N.\nNewman, "Challenges of computational verification in social multimedia," in\nProceedings of the 23rd International Conference on World Wide Web, 2014:\nACM, pp. 743-748.\n[45] C. Boididou, S. Papadopoulos, L. Apostolidis, and Y. Kompatsiaris,\n"Learning to detect misleading content on twitter," in Proceedings of the 2017\nACM on International Conference on Multimedia Retrieval, 2017: ACM, pp.\n278-286.\n[46] O. Russakovsky et al., "Imagenet large scale visual recognition challenge,"\nInternational journal of computer vision, vol. 115, no. 3, pp. 211-252, 2015.\n[47] Y. Kim, "Convolutional neural networks for sentence classification," arXiv\npreprint arXiv:1408.5882, 2014.\n[48] I. Solaiman et al., "Release strategies and the social impacts of language\nmodels," arXiv preprint arXiv:1908.09203, 2019\n[49] J. Baxter, "A Bayesian/information theoretic model of learning to learn via\nmultiple task sampling," Machine learning, vol. 28, no. 1, pp. 7-39, 1997.\n[50] R. Collobert and J. Weston, "A unified architecture for natural language\nprocessing: Deep neural networks with multitask learning," in Proceedings of\nthe 25th international conference on Machine learning, 2008: ACM, pp. 160-\n167.\n[51] L. Derczynski, K. Bontcheva, M. Liakata, R. Procter, G. W. S. Hoi, and A.\nZubiaga, "SemEval-2017 Task 8: RumourEval: Determining rumour veracity\nand support for rumours," arXiv preprint arXiv:1704.05972, 2017.\n[52] R. Caruana, "Multitask learning: A knowledge-based source of inductive bias.\nMachine Learning," 1997.\n[53] S. Ruder, "An overview of multi-task learning in deep neural networks," arXiv\npreprint arXiv:1706.05098, 2017.\n[54] E. Kochkina, M. Liakata, and A. Zubiaga, "All-in-one: Multi-task learning for\nrumour verification," arXiv preprint arXiv:1806.03713, 2018.\n[55] O. Enayet and S. R. El-Beltagy, "NileTMRG at SemEval-2017 Task 8:\nDetermining Rumour and Veracity Support for Rumours on Twitter," in\nProceedings of the 11th International Workshop on Semantic Evaluation\n(SemEval-2017), 2017, pp. 470-474.\n[56] A. Zubiaga, M. Liakata, R. Procter, G. W. S. Hoi, and P. Tolmie, "Analysing\nhow people orient to and spread rumours in social media by looking at\nconversational threads," PloS one, vol. 11, no. 3, p. e0150989, 2016.\n[57] E. Kochkina, M. Liakata, and I. Augenstein, "Turing at semeval-2017 task 8:\nSequential approach to rumour stance classification with branch-lstm," arXiv\npreprint arXiv:1704.07221, 2017.\n[58] T. Mikolov, K. Chen, G. Corrado, and J. Dean, "Efficient estimation of word\nrepresentations in vector space," arXiv preprint arXiv:1301.3781, 2013.\n[59] D. Britz, "Understanding convolutional neural networks for NLP," URL:\nhttp://www. wildml. com/2015/11/understanding-convolutionalneuralnetworks-for-nlp/(visited on 11/07/2015), 2015.\n[60] T. Perry, "Convolutional methods for text," ed: Medium, 2017. zh_TW dc.identifier.doi (DOI) 10.6814/NCCU202000234 en_US
