Dual-Channel and Hierarchical Graph Convolutional Networks for document-level relation extraction

作者：

Highlights：

• A graph convolutional based model is proposed to extract document-level relations.

• The hierarchical graphs model interactive information between entities.

• A dual-channel module supplements low-dimensional contextual information.

• A clinical document-level relation extraction dataset is proposed.

摘要

•A graph convolutional based model is proposed to extract document-level relations.•The hierarchical graphs model interactive information between entities.•A dual-channel module supplements low-dimensional contextual information.•A clinical document-level relation extraction dataset is proposed.

论文关键词：Document-level relation extraction,Graph Convolutional Network,Clinical data

论文评审过程：Received 2 November 2021, Revised 16 April 2022, Accepted 27 May 2022, Available online 3 June 2022, Version of Record 9 June 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117678