Introduction
Event extraction (EE), traditionally modeled as detecting trigger words and extracting corresponding arguments from plain text, plays a vital role in natural language processing since it can produce valuable structured information to facilitate a variety of tasks, such as knowledge base construction, question answering, language understanding, etc.事件抽取( Event Extraction,EE )传统上被建模为从纯文本中检测触发词并提取相应的论元,它在自然语言处理中发挥着至关重要的作用,因为它可以产生有价值的结构化信息,促进各种任务,如知识库构建、问答、语言理解等。
have two critical challenges to EE:
- arguments of one event record may scatter across multiple sentences of the document
- document is likely to contain multiple such event records
However, the sequence tagging model for SEE cannot handle multi-event sentences elegantly, and even worse, the context-agnostic argumentscompletion strategy fails to address the argumentsscattering challenge effectively.
contributions:
- We propose a novel model, Doc2EDAG, which can directly generate event tables based on a document, to address unique challenges of DEE effectively.提出了一种新的基于文档直接生成事件表的模型Doc2EDAG,有效地解决了DEE的独特挑战
- We reformalize a DEE task without trigger words to ease the DS-based document-level event labeling.改造了一个无触发词的DEE任务,简化了基于DS的文档级事件标注
- We build a large-scale real-world dataset for DEE with the unique challenges of arguments-scattering and multi-event, the extensive experiments on which demonstrate the superiority of Doc2EDAG.构建了一个大规模的DEE真实数据集,该数据集具有参数分散和多事件的独特挑战,大量实验证明了Doc2EDAG的优越性
several key notions:
- entity mention: an entity mention is a text span that refers to an entity object
- event role: an event role corresponds to a predefined field of the event table
- event argument: an event argument is an entity that plays a specific event role
- event record: an event record corresponds to an entry of the event table and contains several arguments with required roles