云南省人工智能重点实验室

科研成果

当前位置：首页>科研成果>论文列表

论文作者：	Qingmeng Hu; Hongbin Wang
论文名称：	Chinese Event Extraction Based on Hierarchical Attention Mechanism
论文发表刊物：	17th CCF Conference on Computer Supported Cooperative Work and Social Computing, ChineseCSCW 2022
会议地点：
会议时间：
卷号页码：
论文描述：
收录情况：	EI Indexed
论文摘要：
Event extraction try to extract structured event information from unstructured text. Its major researches can be summarized into three categories: classification methods based on feature learning, methods based on question answering and methods based on Seq2seq. All of these methods are difficult to deal with multi-event sentence. In order to solve this problem, we use hierarchical attention mechanism to treat event extraction as a relation classification task. The BERT and CRF are first used to identify candidate triggers and arguments, and then the hierarchical attention mechanism is used to identify the relationship between trigger and argument. Experiments on both ACE2005 and CEC show that our method outperforms in both trigger classification and argument classification
点击此处下载文章内容