Knowledge graph reasoning relation prediction method based on graph neural network

Abstract

The invention provides a knowledge graph relation reasoning method based on a graph neural network, and the method comprises the steps: mainly dividing a relation reasoning model in knowledge graph reasoning into a scoring function and a prediction result, wherein a candidate result set composed of the target node and the relationship can be obtained according to different modes according to the result of the relationship reasoning. In the process of reasoning the correct relationship in the candidate result set, the method mainly combines the structure information in the knowledge graph with the semantic and attention mechanism of the knowledge graph to realize the reasoning prediction of the unknown relationship in the knowledge graph. In the process of obtaining semantic information of the knowledge graph, structural information around target head and tail entities is extracted by using an isomorphic information theory, attention information collection is performed on the knowledge graph around a target relationship, and a candidate result with the highest score is output as a final result through effective fusion of an attention mechanism.

Description

technical field [0001] The present invention relates to the field of knowledge map reasoning and completion, in particular to a method for predicting knowledge map reasoning relations based on a graph neural network. Background technique [0002] The knowledge graph is a graph structure composed of entities and relationships, which itself has the characteristics of graph nodes and edges. Since the knowledge graph is the record and statistics of real things and relationships, the knowledge graph is inherently an incomplete graph (does not exist in reality. There is a direct relationship between everything, and not all relationships already exist in the knowledge graph). Complementary reasoning for the knowledge graph of an incomplete graph is naturally performed on the possible edges and possible points in the graph. [0003] In the existing knowledge map reasoning methods, a large number of knowledge embedded reasoning methods are used; the advantage of this reasoning metho...

AI Technical Summary

Problems solved by technology

[0005] However, algorithms based on neural networks have problems with their uninterpretable principles, and in the information acquisition of knowledge graphs, due to the existence of convolution algorithms, yes, information can only be acquired mechanically and by obtaining the corresponding receptive field entity data of the convolution center. information

Due to the particularity of the knowledge graph, the complex structural relationship between the target and the surrounding entities will lose information in the convolution operation

Therefore, there is a problem of only accepting fixed-range information and loss of characteristic information

Images