Remote supervision relation extraction method and device based on consistency text enhancement

Abstract

The invention provides a remote supervision relation extraction method and device based on consistency text enhancement, and the method comprises the steps: dividing a plurality of sentence instances according to entity pairs and relation labels, and obtaining a plurality of sentence packages; adopting different text enhancement methods for each sentence instance in each sentence packet, and obtaining a strong enhancement sample and a weak enhancement sample corresponding to each sentence instance in each sentence packet; determining a noise sample, and training the relation prediction model through the unrelated sentence instance and the strong enhancement sample and the weak enhancement sample of the noise sample to obtain a trained relation prediction model; and predicting a to-be-predicted sentence package by using the trained relationship prediction model to obtain a relationship label corresponding to the to-be-predicted sentence package. According to the method, the scale of a data set can be increased through consistent text enhancement, the generalization learning ability of the model can be enhanced, and the NA category and noise sample constraint model can learn more supervision information.

Description

technical field [0001] The invention relates to the field of natural language processing, in particular to a method and device for extracting remote supervision relations based on consistent text enhancement. Background technique [0002] Massive information on the Internet can extract a lot of valuable knowledge and information through related technologies of information extraction. As an important link in information extraction, Relation Extraction (Relation Extraction, RE) aims to extract the relationship between entities from text, and provide other natural language applications such as building knowledge graphs, search engines, dialogue generation, natural question answering, information retrieval, etc. provided important support. [0003] The training of relation extraction models requires a large number of labeled samples to provide supervision information. However, the same relation type may have different text expressions, and at the same time, relations of differ...

AI Technical Summary

Problems solved by technology

[0008] 2) Many research methods reduce the weight of noise samples in the training set or directly filter them out, so that the effective information contained in these noise samples cannot be fully utilized;

[0009] 3) The disturbance added by methods such as confrontation generation, although it can increase the anti-disturbance ability of the model, it usually cannot provide disturbances that meet the actual situation, is not stable, and tends to deviate the direction of model training

Images