Entity unification method for establishing entity unified model based on knowledge graph

Abstract

The invention discloses a method for establishing an entity unified model based on a knowledge graph. The method comprises the following steps: constructing first training data of an entity based on description information and second training data based on structural information; respectively obtaining structure representation vectors of the head entity, the relation and the tail entity by adopting the TransH model; adopting a BERT + Bi-LSTM model to encode the description information of the head entity and the tail entity to obtain a description representation vector, and connecting a CRF layer behind the BERT + Bi-LSTM model to carry out named entity recognition on codes; combining the structure representation vector with the description representation vector to carry out triple verification on the BERT + Bi-LSTM model training; and taking the BERT + Bi-LSTM model after the training as an entity unified model.

Description

technical field [0001] The present application relates to the field of knowledge graphs, in particular to an entity unified model and entity unified method based on knowledge graphs. Background technique [0002] A knowledge graph is essentially a knowledge base of a semantic network, a graph-based data structure consisting of nodes and edges. In recent years, the construction of domain-specific knowledge graphs has also received more and more attention. Oriented to a specific field, build an industry knowledge map based on industry data and relevant knowledge provided by experts in the field, connect different types of information to obtain a relationship network between entities, and provide a specific field from the perspective of "association relationship" to analyze way of reasoning about the problem. [0003] At present, due to the strong professionalism of the chemical industry, the application of knowledge graphs in the chemical industry is still relatively rare. A...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved in this application is to provide an entity unification model and entity unification method based on the knowledge map, so as to solve the problem that the existing knowledge representation model can only convert the known entities and relationships into vectors, and the information generated by information extraction The technical problem that new entities are difficult to convert into vectors through existing knowledge representation models

Images