Fig. 1From: Child-Sum EATree-LSTMs: enhanced attentive Child-Sum Tree-LSTMs for biomedical event extractionBiomedical event annotation and visualization. Entity name recognition is the first step. T1 and T2 denote two named entities “TNF/BHA”. The next step is to detect trigger words. T3 and T4 denote two trigger words “Secretion/abolished”, whose event types are Location and Neg_regulaton. The number 1 and 9 denote the trigger word stimulates’s start and end positions, respectively. Likewise, the number 47 and 60 denote the trigger word abolished’s start and end positions, respectively. E denotes the event type, which is decided by the trigger words. Theme and Cause refer to the event argument, which expresses the event’s semantic rolesBack to article page