Design and Implementation of Text Understanding System Based on Semantic Tagging Instances

Abstract

There are many challenges in text semantic understanding research, such as sparse semantic features, massive representation forms, complex semantic redundancy, diverse echo phenomena, etc. Based on explicit representation of human thinking responses, the importance of semantic tagging and the principles of tagging methods were emphasized in this paper; The routes of large-scale semantic tagging and text semantic understanding system were designed; The concept of semantic scene was defined; By the theory of semantic scene segmentation and recognition, the default information of the semantic scene context was supplemented; Taking elementary math word problems as examples, the prototype system of semantic tagging and understanding was implemented by integrating common sense knowledge.