远监督关系抽取去噪研究综述

Research on Denoising of Distent Supervision Relation Extraction

现有的关系抽取方法大多数为有监督学习方法,需要大量的标注数据,远监督方法解决该问题,可以高效地标注大量的训练数据,使得关系抽取任务得以快速发展。但是,远监督标注的数据集中存在大量的噪声数据,这些噪声数据会影响关系抽取模型的效果。为了避免噪声影响,如何去除选监督数据集中的噪声成为近年来关系抽取任务的一个研究热点。介绍远监督关系抽取去噪的研究历程、方法和相关数据集。

Most of the existing relation extraction methods are supervised learning methods,which require a large amount of labeled data.Distent su pervision method solves this problem and can efficiently label a large amount of training data,so that the relation extraction can develop rapidly.However,there is a large amount of noise data in the distent supervision labeled data set,which will affect the effect of the relation extraction model.In order to avoid the influence of noise,how to remove the noise in the selected supervised data set has become a research hotspot in the relation extraction in recent years.This paper mainly introduces the research process,methods and related data sets on de noising of distent supervised relation extraction.