Entity alignment method, device and system for multi-source knowledge graph fusion

Abstract

The invention discloses an entity alignment method, device and system for multi-source knowledge graph fusion, and relates to the technical field of big data processing. The scheme comprises the following steps: extracting entity features of entities in a knowledge graph, generating an entity embedding matrix according to the entity features of the entities, and obtaining entity representation of the knowledge graph according to the entity embedding matrix; calculating relation information between the entity and the adjacent entity according to the entity representation, and enhancing the entity representation according to the relation information to obtain a complete entity representation; obtaining a final entity embedding matrix according to the complete entity representation; calculating a loss function according to the final entity embedding matrix and the data set; and processing the attribute information of the loss function and the entity by adopting a bidirectional global filtering strategy to generate an iterative positive sample set and an iterative negative sample set, and carrying out iterative training on the neural network model through the sample sets. According to the scheme, the technical problems of insufficient influence interaction and low-quality bootstrap between the entity and the relationship in the prior art are solved.

Description

technical field [0001] The present invention relates to the technical field of big data processing, in particular to an entity alignment method, device and system for multi-source knowledge graph fusion. Background technique [0002] Knowledge Graph (KG for short), consisting of points (entities) and edges (relationships between entities, entity attributes), plays a pivotal role in many researches and applications of artificial intelligence. It is used as a question answering and recommendation system The cornerstone of technology in other fields has received extensive attention. Widely used in knowledge-driven AI tasks, such as question answering models, recommendation systems, search engines, and more. General knowledge graphs and domain knowledge graphs are constructed by different organizations, experts, or automated and semi-automated systems, and there are overlaps and intersections of knowledge between them. It has special significance to promote downstream tasks an...

AI Technical Summary

Problems solved by technology

However, the existing work focuses on the connectivity of the graph, ignoring the relationship type, relationship direction, contribution of entity information to the relationship representation, etc.

[0005] (2) Low quality bootstrapping

The bootstrap method proposed to solve the shortcoming of the lack of pre-aligned seed data believes that if the model is confident in its predicted results, then the results should be considered correct and added to the model training as additional data, so that the model effect can be improved. Both BootEA and MRAEA are excellent and classic bootstrapping methods, but they all rely heavily on the effect of the model itself, and the generated data has a high error rate and low quality, and can only generate positive examples and cannot generate negative examples, which leads to the prediction The problem with low utilization as a result

Images