Few-shot Knowledge Graph Completion Method Based on Meta-learning

A technology of knowledge graph and meta-learning, applied in neural learning methods, database models, biological neural network models, etc., can solve problems affecting the effect of knowledge graph and long-tail relationship, and achieve good robustness and reliability High, good effect

Active Publication Date: 2022-04-29
STATE GRID HUNAN ELECTRIC POWER +2
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Therefore, the embedding method does not work well for long-tail relationships, which seriously affects the effect of knowledge graph completion.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Few-shot Knowledge Graph Completion Method Based on Meta-learning
  • Few-shot Knowledge Graph Completion Method Based on Meta-learning
  • Few-shot Knowledge Graph Completion Method Based on Meta-learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0049] Such as figure 1 Shown is a schematic flow chart of the method of the present invention: the meta-learning-based few-sample knowledge graph completion method provided by the present invention includes the following steps:

[0050] S1. Obtain the knowledge graph to be completed and the corresponding neighborhood knowledge graph; the neighborhood knowledge graph includes neighborhood information of all entities in the knowledge graph to be completed;

[0051] S2. Use the neighborhood knowledge graph obtained in step S1 to initialize the entity embedding in the knowledge graph to be completed; specifically, use the embedding method to train on the neighborhood knowledge graph obtained in step S1 to obtain the knowledge graph to be completed The embedded representation of the entity; and if the neighborhood knowledge graph does not exist, randomly initialize the embedded representation of the entity of the knowledge graph to be completed;

[0052] S3. Divide the relationsh...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a few-sample knowledge map completion method based on meta-learning, which includes acquiring the knowledge map to be completed and the corresponding neighborhood knowledge map; initializing the entity embedding in the knowledge map to be completed; The relational set of the complete knowledge map is divided into training relational set and test relational set, and a triple corresponding to a relation is selected to build a meta-training task or a meta-testing task; training on several batches of meta-training tasks results in few-sample Knowledge graph completion model; use the trained few-sample knowledge graph completion model to complete the meta-test task. The method of the present invention can complete the knowledge map based on a small number of triples, and solve the problem that the traditional embedding-based method has a poor completion effect on the few-sample relationships in the knowledge map, and is more dependent on the neighborhood knowledge map. Low, better robustness, higher reliability, better effect.

Description

technical field [0001] The invention belongs to the field of machine learning, and in particular relates to a meta-learning-based few-sample knowledge graph completion method. Background technique [0002] A knowledge graph is a type of multi-source data that contains various types of relationships and entities. Knowledge graphs are widely used in question answering systems, search engines, recommendation systems and other fields. The knowledge graph is composed of a large number of triples, each of which consists of a head entity, a relation and a tail entity, representing knowledge in the real world. [0003] Although there are a large number of entities, relations, and triples in knowledge graphs, knowledge graphs are generally incomplete, so they need to be completed. By capturing the connections between different types of relationships and entities in the knowledge graph and aggregating features from multi-source data, the knowledge graph is automatically completed, w...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06F16/28G06N3/04G06N3/08
CPCG06F16/288G06N3/08G06N3/045
Inventor 向行陈毅波蒋志怡黄鑫蒋破荒田建伟朱宏宇祝视吕欣琪高建良
Owner STATE GRID HUNAN ELECTRIC POWER
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap