TP-DDI: Transformer-based pipeline for the extraction of Drug-Drug Interactions

作者：

Highlights：

• This research study proposed an end-to-end Drug-Drug Interaction pipeline

• The proposed approach out-performs current state-of-the-art end-to-end systems

• A set of rules is used to mitigate error propagation in the pipeline

• The proposed approach is based on the Transformer architecture

摘要

•This research study proposed an end-to-end Drug-Drug Interaction pipeline•The proposed approach out-performs current state-of-the-art end-to-end systems•A set of rules is used to mitigate error propagation in the pipeline•The proposed approach is based on the Transformer architecture

论文关键词：Drug-drug interaction,Relationship extraction,Drug named entity recognition,Relation classification,Pipeline

论文评审过程：Received 11 January 2021, Revised 17 July 2021, Accepted 16 August 2021, Available online 23 August 2021, Version of Record 25 August 2021.

论文官网地址：https://doi.org/10.1016/j.artmed.2021.102153