Fig. 1From: BioByGANS: biomedical named entity recognition by fusing contextual and syntactic features through graph attention network in node classification frameworkAn example of using the graph to model the topology of a sentence. The topology of a sentence is a graph instead of a sequence. As shown in the figure, challenge is neighbor of recurred in the sentence graph, though challenge is far away from recurred in the sequence. There exists 2 biomedical NEs in the sentence, which serve as subject and object structure respectively. The sentence graph reflects the correlation path and indicates the syntactic relationship between the 2 entities, which helps to implement the NER taskBack to article page