Causal relationship extraction method based on BERT semantic enhancement

Abstract

The invention discloses a causal relationship extraction method based on BERT semantic enhancement. The causal relationship extraction method comprises the steps of causal relationship candidate word library, BERT pre-training and causal relationship extraction. The method is an information extraction technology for rapidly extracting the causal relationship in the text, and the core task is to establish a basic model and an enhancement model under the architecture of a LeakGAN adversarial neural network model to carry out adversarial learning to obtain high-distinction-degree features, analyze the causal relationship in the comment text, and realize deep extraction under semantic enhancement. The method improves the accuracy of causal relationship extraction based on the characteristic that the antagonism learning of the adversarial neural network has more distinction degree, and can be applied to the aspects of event prediction, question and answer systems, scene generation and the like.

Description

technical field [0001] The invention relates to the field of causal relationship extraction, in particular to a causal relationship extraction method based on BERT semantic enhancement. Background technique [0002] In recent years, causality extraction techniques have had an impact on all aspects of natural language processing tasks and have been widely used. Due to the uniqueness and diversity of causal relationship patterns in different fields, the complexity of the semantic structure of review texts, and the diversity of expression methods, it is inevitable to increase the difficulty of causal relationship extraction from review texts in different fields. [0003] In event prediction, question answering system and scenario generation, causality extraction technology has high application value. At the same time, due to the large amount of redundant review text information, it is necessary to use machine learning methods to quickly extract valuable information, so causali...

AI Technical Summary

Problems solved by technology

This is a defect that is difficult to learn semantic fuzzy features in the existing causal relationship extraction. Combining BERT pre-training technology and LeakGAN against the neural network model, a semantically enhanced causal relationship extraction method is proposed.

Images