Aiming at the problem that existing social media fake news detection methods overlook the deep semantic dependency structure in posts, a knowledge-enhanced heterogeneous graph attention fake news detection method is proposed. Firstly, through heterogeneous textual graphs, the intrinsic semantic dependencies and extrinsic knowledge associations of different semantic units in the posts are depicted in detail. Secondly, a dual-layer graph attention mechanism is designed to learn the pattern features of the posts and the knowledge-enhanced entity semantic representations, thereby enhancing the model’s fine-grained semantic perception of the post content and its ability to capture fake news clues from multiple perspectives. The experimental results on two public datasets show that the overall detection performance of this method is superior to the existing text-based content detection methods at present, effectively enhancing the early detection capability of false information on social media.
设 A 表示文本异质图的邻接矩阵。对于任意类型的节点u,其对应不同类型的节点v的之间的邻接关系表示为
式中,。
对于异质文本图不同类型节点初始特征表示,使用BERT(Bidirectional Encoder Representations from Transformers)预训练模型进行初始化:、、。表示所有实体类型词的初始特征表示;表示所有模式类型词的初始特征表示;表示所有概念描述的初始特征表示;n表示帖文中包含的实体的数量;l表示帖文中包含的模式词的数量。由、和堆叠而成,表示图中所有节点组成的初始节点特征表示矩阵,其中,。
ZHANGX Y, CAOJ, LIX R, et al. Mining dual emotion for fake news detection[C]∥Proceedings of the Web Conference 2021. New York, USA: ACM, 2021:3465-3476.
[2]
HUB Z, SHENGQ, CAOJ, et al. Bad actor, good advisor: exploring the role of large language models in fake news detection[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2024,38(20):22105-22113.
[3]
PANL M, WUX B, LUX Y, et al. Fact-checking complex claims with program-guided reasoning[C]∥Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics. Stroudsburg, USA: ACL, 2023:6981-7004.
[4]
ZHANGX, GAOW. Towards LLM-based fact verification on news claims with a hierarchical step-by-step prompting method[C]∥Proceedings of the 13th International Joint Conference on Natural Language Processing and the 3rd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics. Stroudsburg, USA: ACL, 2023:996-1011.
[5]
WANGY Z, QIANS S, HUJ, et al. Fake news detection via knowledge-driven multimodal graph convolutional networks[C]∥Proceedings of the 2020 International Conference on Multimedia Retrieval. New York, USA: ACM, 2020:540-547.
[6]
HUL M, YANGT C, ZHANGL H, et al. Compare to the knowledge: graph neural fake news detection with external knowledge[C]∥Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing. Stroudsburg, USA: ACL, 2021:754-763.
XIAOL, ZHANGQ, SHIC Y, et al. MSynFD: multi-hop syntax aware fake news detection[C]∥Proceedings of the ACM Web Conference 2024. New York, USA: ACM, 2024:4128-4137.
[9]
LUOY J, RUX Y, LIUK W, et al. OneKE: a dockerized schema-guided LLM agent-based knowledge extraction system[DB/OL]. (2024-12-28)[2025-03-12].
[10]
HEH, CHOIJ D. The stem cell hypothesis: dilemma behind multi-task learning with transformer encoders[C]∥Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. Stroudsburg, USA: ACL, 2021:5555-5577.
[11]
SHENGQ, ZHANGX Y, CAOJ, et al. Integrating pattern- and fact-based fake news detection via model preference learning[C]∥Proceedings of the 30th ACM International Conference on Information & Knowledge Management. New York, USA: ACM, 2021:1640-1650.
[12]
DUZ X, QIANY J, LIUX, et al. GLM: general language model pretraining with autoregressive blank infilling[C]∥Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics. Stroudsburg, USA: ACL, 2022:320-335.
[13]
HUE J, SHENY, WALLISP, et al. Lora: Low-rank adaptation of large language models[DB/OL]. (2021-10-16)[2024-08-24].