Login / Signup

Document-Level Biomedical Relation Extraction Using Graph Convolutional Network and Multihead Attention: Algorithm Development and Validation.

Jian WangXiaoyu ChenYu ZhangYijia ZhangJiabin WenHongfei LinZhihao YangXin Wang
Published in: JMIR medical informatics (2020)
The GCN model can effectively exploit the across sentence dependency information to improve the performance of intersentence CDR extraction. Both the deep context representation and multihead attention are helpful in the CDR extraction task.
Keyphrases
  • neural network
  • working memory
  • machine learning
  • deep learning
  • healthcare