Semantic Inheritance and Overloading
Ping Zhu, Pohua Lv, Jin Shi, Xuetao Jiang, Weiming Zou, Yirong Ma · 2022
Humanoid reasoning and computing are the crucial research objectives of computing thinking, so semantic understanding and explicit representation are the basis steps for computing thinking. However, at present, there is no mature and systematic theory and engineering technology to fully describe massive multi-dimensional semantic expressions of natural language. Then the semantic inheritance and overloading theory was proposed to describe the semantic superposition and synthesis processing of the input vocabularies in this paper. Taking the clause as the basic semantic unit, the semantics were integrated by the local semantics of the clause itself and the global semantics of the cross clauses. The semantic frameworks were represented by the core vocabulary sequence of the clause, and the unified semantic representation grammar of the semantic frameworks to explicitly describe the local semantics and global semantics was studied; From the perspective of big data engineering, according to the unified grammar, the semantic annotation was chose as the initial step, the semantic scene based on the pattern matching of semantic frameworks was divided, the implementation technologies of semantic inheritance and overloading theory between scenes on the base of the observed language representation phenomenon were put forward, and the core algorithms of machine humanoid resolving primary mathematics application problems were realized, the effectiveness of semantic inheritance and overloading theory was preliminarily verified. The advantage was that the implementation could make full use of the power of the cloud computing to do semantic annotation and semantic framework pattern matching in parallel, and explore the engineering way to achieve natural language semantic understanding and computing thinking.