Relation Extraction

A relationship extraction task requires the detection and classification of semantic relationship mentions within a set of artifacts, typically from text or XML documents. (Understanding Conference in 1998/ binary relations)
Approaches:

1. text-based relationship extraction: rely on the use of pretrained relationship structure
2. use of domain ontologies
3. involves visual detection of meaningful relationships in parametric values of objects listed on a data table that shift positions as the table is permuted
   (http://en.wikipedia.org/wiki/Relationship_extraction)

---

Methods:
There are five different methods of doing Relation Extraction:
进行关系提取有五种不同的方法：

1. Rule-based RE 基于规则的资源检索 - through hand-crafted patterns

• Pro:
  • Humans can create pattern which tend to have high precisio
  • Can be tailored to specific domains
• Cons:
  • A lot of manual work to create all possible rules
  • Have to create rules for every relation type

2. Weakly Supervised RE 弱监督 RE - start out with a set of hand-crafted rules and automatically find new ones from the unlabeled text data, through and iterative process (bootstrapping).
3. Supervised RE 监督式 RE
4. Distantly Supervised RE 远程监督关系推理
5. Unsupervised RE 无监督 RE
   (https://medium.com/@andreasherman/different-ways-of-doing-relation-extraction-from-text-7362b4c3169e)

---

model:(https://nlpprogress.com/english/relationship_extraction.html)
6. End-to-end model: It uses external lexical resources, such as WordNet, part-of-speech tags, dependency tags, and named entity tags.

Model	F1	Paper / Source	Code
BERT-based Models
A-GCN (Tian et al., 2021)	89.85	Dependency-driven Relation Extraction with Attentive Graph Convolutional Networks	Official
Matching-the-Blanks (Baldini Soares et al., 2019)	89.5	Matching the Blanks: Distributional Similarity for Relation Learning
R-BERT (Wu et al. 2019)	89.25	Enriching Pre-trained Language Model with Entity Information for Relation Classification	mickeystroller’s Reimplementation
CNN-based Models
Multi-Attention CNN (Wang et al. 2016)	88.0	Relation Classification via Multi-Level Attention CNNs	lawlietAi’s Reimplementation
Attention CNN (Huang and Y Shen, 2016)	84.3
85.9*	Attention-Based Convolutional Neural Network for Semantic Relation Extraction
CR-CNN (dos Santos et al., 2015)	84.1	Classifying Relations by Ranking with Convolutional Neural Network	pratapbhanu’s Reimplementation
CNN (Zeng et al., 2014)	82.7	Relation Classification via Convolutional Deep Neural Network	roomylee’s Reimplementation
RNN-based Models
Entity Attention Bi-LSTM (Lee et al., 2019)	85.2	Semantic Relation Classification via Bidirectional LSTM Networks with Entity-aware Attention using Latent Entity Typing	Official
Hierarchical Attention Bi-LSTM (Xiao and C Liu, 2016)	84.3	Semantic Relation Classification via Hierarchical Recurrent Neural Network with Attention
Attention Bi-LSTM (Zhou et al., 2016)	84.0	Attention-Based Bidirectional Long Short-Term Memory Networks for Relation Classification	SeoSangwoo’s Reimplementation
Bi-LSTM (Zhang et al., 2015)	82.7
84.3*	Bidirectional long short-term memory networks for relation classification

7. Dependency Models

Model	F1	Paper / Source 论文/来源	Code 代码
BRCNN (Cai et al., 2016) BRCNN（Cai 等人，2016）	86.3	Bidirectional Recurrent Convolutional Neural Network for Relation Classification 用于关系分类的双向循环卷积神经网络
DRNNs (Xu et al., 2016) DRNN（Xu 等人，2016 年）	86.1	Improved Relation Classification by Deep Recurrent Neural Networks with Data Augmentation 通过数据增强的深度循环神经网络改进关系分类
depLCNN + NS (Xu et al., 2015a) depLCNN + NS（徐等人，2015a）	85.6	Semantic Relation Classification via Convolutional Neural Networks with Simple Negative Sampling 通过简单负采样的卷积神经网络进行语义关系分类
SDP-LSTM (Xu et al., 2015b) SDP-LSTM（徐等人，2015b）	83.7	Classifying Relations via Long Short Term Memory Networks along Shortest Dependency Path 通过长短期记忆网络沿最短依赖路径对关系进行分类	Sshanu’s Reimplementation Sshanu 的重新实现"
"DepNN (Liu et al., 2015) DepNN（Liu 等人，2015）	83.6	A Dependency-Based Neural Network for Relation Classification 基于依赖关系的神经网络进行关系分类
FCN (Yu et al., 2014) FCN（Yu 等人，2014）	83	Factor-based compositional embedding models 基于因子的组合嵌入模型
MVRNN (Socher et al., 2012) MVRNN（Socher 等人，2012）	82.4	Semantic Compositionality through Recursive Matrix-Vector Spaces 通过递归矩阵向量空间实现语义组合	pratapbhanu’s Reimplementation pratapbhanu 的重新实施

8. New York Times Corpu

Model	P@10% 磷@10%	P@30% 磷@30%	Paper / Source 论文/来源	Code 代码
KGPOOL (Nadgeri et al., 2021) KGPOOL（Nadgeri 等人，2021 年）	92.3	86.7	KGPool: Dynamic Knowledge Graph Context Selection for Relation Extraction KGPool：用于关系提取的动态知识图谱上下文选择	KGPOOL
RECON (Bastos et al., 2021) RECON（Bastos 等人，2021 年）	87.5	74.1	RECON: Relation Extraction using Knowledge Graph Context in a Graph Neural Network RECON：使用图神经网络中的知识图谱上下文进行关系提取	RECON
HRERE (Xu et al., 2019) HRERE（Xu 等，2019）	84.9	72.8	Connecting Language and Knowledge with Heterogeneous Representations for Neural Relation Extraction 将语言和知识与异构表示连接起来以进行神经关系提取	HRERE
PCNN+noise_convert+cond_opt (Wu et al., 2019) PCNN+noise_convert+cond_opt（Wu 等人，2019 年）	81.7	61.8	Improving Distantly Supervised Relation Extraction with Neural Noise Converter and Conditional Optimal Selector 使用神经噪声转换器和条件最优选择器改进远程监督关系提取
Intra- and Inter-Bag (Ye and Ling, 2019) 袋内和袋间 (Ye and Ling，2019)	78.9	62.4	Distant Supervision Relation Extraction with Intra-Bag and Inter-Bag Attentions 利用袋内和袋间注意力机制进行远程监督关系提取	Code 代码
RESIDE (Vashishth et al., 2018) 居住 (Vashishth 等人，2018)	73.6	59.5	RESIDE: Improving Distantly-Supervised Neural Relation Extraction using Side Information RESIDE：利用辅助信息改进远程监督神经关系提取	RESIDE
PCNN+ATT (Lin et al., 2016) PCNN+ATT（Lin 等人，2016 年）	69.4	51.8	Neural Relation Extraction with Selective Attention over Instances 通过对实例的选择性注意进行神经关系提取	OpenNRE 开放NRE
MIML-RE (Surdeneau et al., 2012) MIML-RE（Surdeneau 等人，2012 年）	60.7+	-	Multi-instance Multi-label Learning for Relation Extraction 用于关系提取的多示例多标签学习	Mimlre 我的名字是
MultiR (Hoffman et al., 2011) MultiR（Hoffman 等人，2011 年）	60.9+	-	Knowledge-Based Weak Supervision for Information Extraction of Overlapping Relations 基于知识的弱监督重叠关系信息提取	MultiR 多R
(Mintz et al., 2009) （Mintz 等人，2009 年）	39.9+	-	Distant supervision for relation extraction without labeled data 无需标记数据即可进行远程监督关系提取