基于深度学习的实体关系抽取方法文献综述

易 战军, 曾 锐
宜春学院人工智能与信息工程学院

摘要

实体关系抽取作为自然语言处理（NLP）领域信息抽取的核心任务，在知识图谱构建、智能问答等场景中发挥关
键作用，其效率与准确性直接影响海量非结构化文本的价值挖掘。深度学习凭借自主特征学习能力成为实体关系抽取主流
技术，递归神经网络（RNN）及其变体（LSTM、Bi-LSTM）通过时序建模解决长距离依赖问题，但存在参数训练复杂或
训练周期长等不足；卷积神经网络（CNN）利用卷积与池化操作高效提取局部语义特征，结合位置特征等改进进一步提升
性能，却在长句建模中存在局限；图神经网络（GNN）通过图结构与节点交互优化全局关系建模，在文档级实体关系抽取
中优势显著，其优化聚焦依存树剪枝、多模态图表示等方向；联合学习通过参数共享或序列标注实现端到端建模，部分方
法针对特定领域优化特征融合，提升抽取鲁棒性；大语言模型（LLM）借助大规模预训练与微调增强语义理解和泛化能力，
在低资源、零样本场景及专业术语处理中表现突出，参数高效微调等技术进一步优化其性能。

关键词

实体关系抽取；RNN；LSTM；CNN；GNN；LLM

参考

[1]Socher R,Huval B,Manning C D,et al.Semantic

c o m p o s i t i o n a l i t y t h r o u g h r e c u r s i v e m a t r i x - v e c t o r

spaces[C]//Proc of Empirical Methods in Natural Language

Processing.2012:1201-1211．

[2] NINGTHOUJAM D,YADAV S,BHATTACHARYYA P,et

al.Relation extraction between the clinical entities based on the

shortest dependency path based lstm[J].arXiv:1903.09941,2019.

[3]Liu Chunyang,Sun Wenbo,Chao Wenlan,et al.Convolution

neural network for relation extraction[C]//Advanced Data Mining

and Applications.2013:231-242．

[4]ZENG D,LIU K,LAI S,et al.Relation classification via

convolutional deep neural network[C]//Proceedings of the25th

International Conference on Computational Linguistics:Technical

Papers,2014:2335-2344.

[ 5 ] N G U Y E N T H , G R I S H M A N R . R e l a t i o n

extraction:perspective from convolutional neural networks[C]//

Proceedings of the 1st Workshop on Vector Space Modeling for

Natural Language Processing,2015:39-48.

[6] SONG L,ZHANG Y,GILDEA D,et al.Leveraging

dependency forest for neural medical relation extraction[J].

arXiv:1911.04123,2019.

[7] FU T J,LI P H,MA W Y.GraphRel:modeling text as

relational graphs for joint entity and relation extraction[C]//

Proceedings of the 57th Annual Meeting of the Association for

Computational Linguistics,2019:1409-1418.

[8]Miwa M,Bansal M.End-to-end relation extraction using

LSTMs on sequences and tree structures[C]//Proc of Meeting of the

Association for Computational Linguistics.2016:1105-1116．

[9]Zheng Suncong,Wang Feng,Bao Hongyun,et al.Joint

extraction of entities and relations based on a novel tagging

scheme[C]//Proc of Meeting of the Association for Computational

Linguistics.2017:1227-1236．

[10] 张西硕 , 柳林 , 王海龙 , 等 . 知识图谱中实体关系抽

取方法研究 [J]. 计算机科学与探索 ,2024,18(03):574-596.

[11] WANG J X, ZHANG L L, LEE W S, et al. When

phrases meet probabilities:enabling open relation extraction with

cooperating large language models[C]//Proceedings of the62nd

Annual Meeting of the Association for Computational Linguistics.

Stroudsburg:ACL, 2024:13130-13147.