Introduction

传统上，文本中实体之间的关系抽取问题是被作为两个独立的任务进行研究的：命名实体识别（Named Entity Recognition）和关系抽取（Relation Extraction）。 在过去的几年中，研究者们对实体和关系的联合抽取模型的兴趣激增，本文主要总结一下实体关系联合抽取的各种方法，将这些方法分为以下几种类别（包括传统的Pipelined的方法）。

Pipelined Approach

• A Frustratingly Easy Approach for Entity and Relation Extraction——NAACL 2021

Structured Prediction

• Contrastive Triple Extraction with Generative Transformer——AAAI 2021
• Effective Modeling of Encoder-Decoder Architecture for Joint Entity and Relation Extraction——AAAI 2020
• A Novel Cascade Binary Tagging Framework for Relational Triple Extraction——ACL 2020
• TPLinker: Single-stage Joint Extraction of Entities and Relations Through Token Pair Linking——COLING 2020
• Two are Better than One: Joint Entity and Relation Extraction with Table-Sequence Encoders——EMNLP 2020
• Joint Entity and Relation Extraction with Set Prediction Networks——2020

Multi-task Learning

• Progressive Multitask Learning with Controlled Information Flow for Joint Entity and Relation Extraction——AAAI 2021

标签：抽取,Joint,Entity,Extraction,Relation,AAAI,Summary
来源： https://blog.csdn.net/qq_41329175/article/details/115447774