Explainable document-level event extraction via back-tracing to sentence-level event clues

作者：

Highlights：

• An explainable document-level event extraction framework ExpDEE is proposed.

• ExpDEE can reasonably explain extracted results from relevant texts pieces.

• Experimental results on two public benchmarks show that the superiority of ExpDEE.

摘要

•An explainable document-level event extraction framework ExpDEE is proposed.•ExpDEE can reasonably explain extracted results from relevant texts pieces.•Experimental results on two public benchmarks show that the superiority of ExpDEE.

论文关键词：Document-level event extraction,Explainable framework

论文评审过程：Received 13 September 2021, Revised 1 March 2022, Accepted 30 March 2022, Available online 22 April 2022, Version of Record 13 May 2022.

论文官网地址：https://doi.org/10.1016/j.knosys.2022.108715