文件列表:

KG (0, 2021-09-08)
KG\USPoliticalKG.data (388689, 2021-09-08)
SemEval.py (12875, 2021-09-08)
SemEval_ablation_GNN.py (15605, 2021-09-08)
SemEval_ablation_X.py (13238, 2021-09-08)
SemEval_ablation_edge1.py (13266, 2021-09-08)
SemEval_ablation_edge2.py (13153, 2021-09-08)
SemEval_ablation_entire_graph.py (12881, 2021-09-08)
SemEval_inference.py (12842, 2021-09-08)
SemEval_remove_knowledge.py (13539, 2021-09-08)
crawler.py (220, 2021-09-08)
fold (0, 2021-09-08)
fold\fold_0.txt (306, 2021-09-08)
fold\fold_1.txt (312, 2021-09-08)
fold\fold_2.txt (310, 2021-09-08)
fold\fold_3.txt (315, 2021-09-08)
fold\fold_4.txt (308, 2021-09-08)
fold\fold_5.txt (313, 2021-09-08)
fold\fold_6.txt (305, 2021-09-08)
fold\fold_7.txt (307, 2021-09-08)
fold\fold_8.txt (306, 2021-09-08)
fold\fold_9.txt (313, 2021-09-08)
graph (0, 2021-09-08)
graph\graph1.zip (20603054, 2021-09-08)
graph\graph2.zip (21202205, 2021-09-08)
ground_truth.txt (4277, 2021-09-08)
roberta (0, 2021-09-08)
roberta\article.txt (2288047, 2021-09-08)
roberta\roberta.py (1215, 2021-09-08)
roberta\title_roberta.pt (1982200, 2021-09-08)
roberta\titles.txt (44427, 2021-09-08)

### Knowledge Graph Augmented Political Perspective Detection in News Media This repository serves as a code listing for the paper 'Knowledge Graph Augmented Political Perspective Detection in News Media'. Arxiv version of the paper is available at https://arxiv.org/abs/2108.03861. Work in progress. #### Abstract Identifying political perspective in news media has become an important task due to the rapid growth of political commentary and the increasingly polarized political ideologies. Previous approaches tend to focus on leveraging textual information and leave out the rich social and political context that helps individuals identify stances. To address this limitation, in this paper, we propose a perspective detection method that incorporates external knowledge of real-world politics. Specifically, we construct a political knowledge graph with 1,071 entities and 10,703 triples. We then construct heterogeneous information networks to represent news documents, which jointly model news text and external knowledge. Finally, we apply gated relational graph convolutional networks and conduct political perspective detection as graph-level classification. Extensive experiments demonstrate that our method consistently achieves the best performance and outperforms state-of-the-art methods by at least 5.49%. Ablation studies further bear out the necessity of external knowledge and the effectiveness of our graph-based approach. #### Knowledge Graph We construct a contemporary U.S. political knowledge graph and present it in /KG. It could serve as external knowledge for various tasks such as perspective detection and misinformation detection.

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：