Fig. 3From: Drug knowledge discovery via multi-task learning and pre-trained modelsThe base architecture diagram of our RE modelBack to article page