Recurrent neural networks with segment attention and entity description for relation extraction from clinical texts

作者：

Highlights：

• We introduce concat-attention mechanism to capture the most important words for relation extraction from local and global.

• We propose a segment attention mechanism to improve the performance of model processing long sentences.

• A tensor-based entity description is proposed to improve the performance when there are multiple entities in a sentence.

• Our model improves the F1-score by approximately 3% compared with baseline model.

摘要

•We introduce concat-attention mechanism to capture the most important words for relation extraction from local and global.•We propose a segment attention mechanism to improve the performance of model processing long sentences.•A tensor-based entity description is proposed to improve the performance when there are multiple entities in a sentence.•Our model improves the F1-score by approximately 3% compared with baseline model.

论文关键词：Segment attention mechanism,Tensor-based entity description,Relation extraction,Clinical texts

论文评审过程：Received 11 December 2018, Revised 23 April 2019, Accepted 23 April 2019, Available online 2 May 2019, Version of Record 22 May 2019.

论文官网地址：https://doi.org/10.1016/j.artmed.2019.04.003