Method and system for expressing knowledge graph and text information based on reference sentence

Abstract

The invention discloses a method and a system for expressing a knowledge graph and text information based on a reference sentence, and relates to the technical field of machine learning. The method comprises the following steps: modeling the knowledge graph to obtain an entity vector and a relation vector; obtaining deep semantic information related to the relationship and contained in the plain text, and performing knowledge modeling to obtain a textualized relationship vector; obtaining deep semantic information related to an entity and contained in the plain text, and performing knowledge modeling to obtain a textualized entity vector; and based on the entity vector, the relation vector, the textualization relation vector, the textualization entity vector and the word vector, constructing an optimization parameter, and realizing joint representation of the knowledge graph and the text information. According to the method, textual modeling is carried out on the entities in the knowledge graph by utilizing the reference sentences, and noise reduction is carried out on joint representation of the knowledge graph and the text information, so that the knowledge representation and deduction quality is improved.

Description

technical field [0001] The invention relates to the technical field of machine learning, in particular to a method and system for representing a knowledge map and text information based on an referring sentence. Background technique [0002] Knowledge graph (Knowledge Graph) has attracted the attention of many research fields in recent years because of its ability to effectively model and describe abstract concepts and concrete examples in the real world. The knowledge graph representation learning method can effectively alleviate the above problems by mapping the entities and relationships in the knowledge graph to a low-dimensional vector space, and enhance the knowledge learning ability in knowledge representation, knowledge deduction, knowledge fusion, knowledge completion, etc. More and more studies have found that additional text information can provide rich semantic resources for knowledge graph representation, which plays an important auxiliary role in optimizing kno...

AI Technical Summary

Problems solved by technology

Some models propose a joint representation learning model that maps the entities in the knowledge graph and the words in the text to the same vector space through an alignment mechanism, and another scheme models the contextual information to a certain extent. The above models use the text However, the widespread noise in the text information has greatly weakened the knowledge representation ability of these models. Not all sentences containing an entity are helpful for explaining and modeling the entity.

Images