Abstract
This article describes the structure of a uniform object capable of representing information about relations between separate basic semantico-syntactic structures in a natural language knowledge base.The first part of the article deals with analyzing relations in natural language and defining their common properties. Particular attention is paid to estimating usefulness of these properties in the broader context of components of knowledge base and connections between them.The second part of the article is focused on mapping the obtained properties to an artificial structure suitable for use in a relational database and determining their limitations.The resulting structure allows the basic structure of the knowledge base to be finalized, providing a complete architecture that can be implemented as a software product to test the basic capabilities of the proposed structure and to determine its limits in solving more complicated tasks.
Highlights
Ever since the appearance of the first computers, unceasing attempts have been made to create an artificial intelligence – a more advanced version of the human intelligence, enhanced by the vast network of knowledge nodes and processing units
The key points of the research are the model of individual speech system (ISS) as a model of a human’s speech processor and its integral parts – knowledge base (KB), where the natural language information (NLI) is stored, and linguistic processor (LP), which handles the two-way transformations of KB structured data and NL text data [1, 2]
Let us see how the complete structure of the relation object looks after all additions: (S1) [Rdirection, type] [Rdata] (S2)
Summary
This article focuses on forming a uniform object capable of representing a relation in a natural language knowledge base. The article describes both the functional requirements for such object and its resulting structure. Creation of this object allows to finalize the basic structure of the knowledge base and to implement it as a software product ― both to test its capabilities and to use it as a foundation for following specialized extensions
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have