计算机工程与应用 ›› 2024, Vol. 60 ›› Issue (5): 122-129.DOI: 10.3778/j.issn.1002-8331.2210-0237

• 模式识别与人工智能 • 上一篇    下一篇

文档级事件抽取反向推理模型

纪婉婷,马宇航,鲁闻一,王俊陆,宋宝燕   

  1. 辽宁大学 信息学院,沈阳 110036
  • 出版日期:2024-03-01 发布日期:2024-03-01

Reverse Inference Model for Document-Level Event Extraction

JI Wanting, MA Yuhang, LU Wenyi, WANG Junlu, SONG Baoyan   

  1. College of Information, Liaoning University, Shenyang 110036, China
  • Online:2024-03-01 Published:2024-03-01

摘要: 事件抽取旨在从非结构化文本中检测事件类型并抽取事件要素。现有方法在处理文档级文本时仍存在局限性。这是因为文档级文本可能由多个事件组成,并且构成某一事件的事件要素通常分散在不同句子中。为应对上述挑战,提出了一种文档级事件抽取反向推理模型(reverse inference model for document-level event extraction,RIDEE)。基于无触发词的设计,将文档级事件抽取转化为候选事件要素抽取和事件触发推理两个子任务,并行式抽取事件要素并检测事件类型。此外,设计了一种用于存储历史事件的事件依赖池,使得模型在处理多事件文本时可以充分利用事件之间的依赖关系。公开数据集上的实验结果表明,与现有事件抽取模型相比,RIDEE在进行文档级事件抽取时具有更优的性能。

关键词: 文档级事件抽取, 反向推理, 无触发词, 事件依赖池

Abstract: Event extraction aims to detect event types and extract event arguments from unstructured texts. Existing methods still have limitations when dealing with document-level texts. This is because a document-level text may consist of multiple events, and the event arguments that constitute an event are usually scattered across different sentences. To address the above challenges, this paper proposes a reverse inference model for document-level event extraction (RIDEE). Based on the design without trigger words, RIDEE simplifies the document-level event extraction into two sub-tasks, candidate event argument extraction and event triggering inference, to extract event arguments in parallel and detect event types. In addition, this paper designs an event dependency pool for storing historical events, so that the model can make full use of the dependencies between events when processing the multi-event texts. Experimental results on the public dataset show that RIDEE has better performance in document-level event extraction than the existing event extraction models.

Key words: document-level event extraction, reverse inference, without trigger words, event dependency pool