深度学习元事件抽取研究

Meta-event eextraction based on deep learning

元事件抽取研究是从模式匹配、机器学习到深度学习。基于深度学习元事件抽取是将事件抽取任务转换成分类任务、序列标注任务、机器阅读理解任务。分析了事件抽取中存在的数据、语义、重叠、级联等问题,总结了元事件抽取未来将面临的待解决问题。

The research of meta-event extraction is from pattern matching,machine learning to the current deep learning.The meta-event extraction based on deep learning is to transfer the information extraction into classification,sequence annotation and machine reading comprehension tasks.Here we analyze data,semantics,overlap and cascade issues in the event extraction and give the perspectives.