Natural Language Semantic Representation Method Based on the Scene Framework

Abstract

This paper constructs a standard semantic reference system with a large number of scene semantic representations, then interprets and understands the reference system with low-level semantics, that is, commonsense, and compares the semantic representations of the standard reference system with the input text, forming the premise and foundation of machine analogy thinking. This paper discusses the method of using the scene framework to construct standard semantic reference system, describes the dynamic process of semantic recognition by analogy with standard reference system, describes the application examples of machine thinking oriented semantic representation, finally, summarizes the paper and explains the application prospect of natural language semantic representation method based on the scene framework.