Knowledge graph construction method based on adaptive few-shot relation extraction

Abstract

The invention discloses a knowledge map construction method based on self-adaptive few-sample relationship extraction, the extraction method includes extracting the relationship between entities using an adaptive relationship extraction model, and the construction of the self-adaptive relationship extraction model includes: S100: use a text encoder to Encode the training set instance to generate contextual semantics; S200: Input the support set into the parameter generator to generate initialization softmax parameters; S300: Input the contextual semantics generated in step S100 into the adaptive graph neural network, and use the adaptive graph neural network to The instance is updated; S400: Use a softmax classifier to classify and predict the updated instance to obtain a relationship type. The present invention does not require a large amount of manual labeling data when acquiring relationships, avoids time-consuming and costly costs caused by a large number of manual markings, and can complete the relationship extraction task in a specific field with a small amount of label data in a specific field.

Description

technical field [0001] The invention belongs to the field of natural language processing, and in particular relates to a knowledge map construction method based on adaptive few-sample relation extraction. Background technique [0002] Knowledge graph, also known as scientific knowledge graph, is a series of different graphs showing knowledge development process and structural relationship, using visualization technology to describe knowledge resources and their carriers, mining, analyzing, constructing, drawing and displaying knowledge and the relationship between them interconnected. In the existing technology, the construction of knowledge graphs for general domains is to use the original unstructured text to form a knowledge graph, which mainly includes steps: (1) extracting entities, that is, automatically identifying entities from unstructured texts; (2) extracting Relationship, which is to identify the relationship between entities; (3) Entity linking, which is to log...

AI Technical Summary

Problems solved by technology

However, the shortcoming of remote supervision is that there is a large amount of noise data in the generated data, which still cannot fundamentally solve the long-tail problem of sample distribution.

Images