【Coref】一些资料的收集

扩展阅读：

• https://www.163.com/dy/article/FSKO8D5N0511DPVD.html
• 21 意境级讲解 共指消解的方法
• 共指消解自我总结-初级
• 事件共指消解系列文章——论文笔记
• 指代消解模型浅析

论文统计-ACL

2016

coref

• Which Coreference Evaluation Metric Do You Trust? A Proposal for a Link-based Entity Aware Metric
• Improving Coreference Resolution by Learning Entity-Level Distributed Representations
• Using mention accessibility to improve coreference resolution

pronoun

• Chinese Zero Pronoun Resolution with Deep Neural Networks

2017

coref

• Joint Learning for Event Coreference Resolution
• Don’t understand a measure? Learn it: Structured Prediction for Coreference Resolution optimizing its measures
• Lexical Features in Coreference Resolution: To be Used With Caution

pronoun

• Generating and Exploiting Large-scale Pseudo Training Data for Zero Pronoun Resolution
• Learning attention for historical text normalization by learning to pronounce

2018

coref

• Improving Event Coreference Resolution by Modeling Correlations between Event Coreference Chains and Document Topic Structures
• Entity-Centric Joint Modeling of Japanese Coreference Resolution and Predicate Argument Structure Analysis
• Neural Coreference Resolution with Deep Biaffine Attention by Joint Mention Detection and Mention Clustering
• Neural Cross-Lingual Coreference Resolution And Its Application To Entity Linking
• Investigating Domain-Specific Information for Neural Coreference Resolution on Biomedical Texts

pronoun

• Deep Reinforcement Learning for Chinese Zero Pronoun Resolution
• The Role of Syntax During Pronoun Resolution: Evidence from fMRI

2019

coref

• End-to-end Deep Reinforcement Learning Based Coreference Resolution
• Coreference Resolution with Entity Equalization
• The KnowRef Coreference Corpus: Removing Gender and Number Cues for Difficult Pronominal Anaphora Resolution
• Using Automatically Extracted Minimum Spans to Disentangle Coreference Evaluation from Boundary Detection
• Revisiting Joint Modeling of Cross-document Entity and Event Coreference Resolution
• Semantically Constrained Multilayer Annotation: The Case of Coreference
• Interconnected Question Generation with Coreference Alignment and Conversation Flow Modeling
• On GAP Coreference Resolution Shared Task: Insights from the 3rd Place Solution

pronoun

• Knowledge-aware Pronoun Coreference Resolution
• Gendered Ambiguous Pronoun (GAP) Shared Task at the Gender Bias in NLP Workshop 2019
• Transfer Learning from Pre-trained BERT for Pronoun Resolution
• MSnet: A BERT-based Network for Gendered Pronoun Resolution
• Look Again at the Syntax: Relational Graph Convolutional Network for Gendered Ambiguous Pronoun Resolution
• Fill the GAP: Exploiting BERT for Pronoun Resolution
• Resolving Gendered Ambiguous Pronouns with BERT
• Anonymized BERT: An Augmentation Approach to the Gendered Pronoun Resolution Challenge
• Gendered Pronoun Resolution using BERT and an Extractive Question Answering Formulation
• Gendered Ambiguous Pronouns Shared Task: Boosting Model Confidence by Evidence Pooling

2020

coref

• Toward Gender-Inclusive Coreference Resolution
• CorefQA: Coreference Resolution as Query-based Span Prediction
• Active Learning for Coreference Resolution using Discrete Annotation
• New Insights into Cross-Document Event Coreference: Systematic Comparison and a Simplified Approach

pronoun

• ZPR2: Joint Zero Pronoun Recovery and Resolution using Multi-Task Learning and BERT

2021

coref

• Coreference Reasoning in Machine Reading Comprehension
• End-to-End AMR Corefencence Resolution
• Exploiting Document Structures and Cluster Consistencies for Event Coreference Resolution
• Benchmarking Scalable Methods for Streaming Cross Document Entity Coreference
• Coreference Resolution without Span Representations
• OntoGUM: Evaluating Contextualized SOTA Coreference Resolution on 12 More Genres
• Scaling Within Document Coreference to Long Texts
• Injecting Knowledge Base Information into End-to-End Joint Entity and Relation Extraction and Coreference Resolution
• Annotation and Evaluation of Coreference Resolution in Screenplays
• Cross-document Coreference Resolution over Predicted Mentions

pronoun

• A Joint Model for Dropped Pronoun Recovery and Conversational Discourse Parsing in Chinese Conversational Speech
• Unsupervised Pronoun Resolution via Masked Noun-Phrase Prediction

论文统计-EMNLP

coref

• 2017:End-to-end Neural Coreference Resolution
• 2017: Revisiting Selectional Preferences for Coreference Resolution
• 2017: Event Coreference Resolution by Iteratively Unfolding Inter-dependencies among Events
• 2018: PreCo: A Large-scale Dataset in Preschool Vocabulary for Coreference Resolution
• 2018: Using Linguistic Features to Improve the Generalization Capability of Neural Coreference Resolvers
• 2018: A Probabilistic Annotation Model for Crowdsourcing Coreference
• 2018: Multi-Task Identification of Entities, Relations, and Coreference for Scientific Knowledge Graph Construction
• 2018: Coreference and Coherence in Neural Machine Translation: A Study Using Oracle Experiments
• 2019: Rewarding Coreference Resolvers for Being Consistent with World Knowledge or v1
• 2019: A Regularization Approach for Incorporating Event Knowledge and Coreference Relations into Neural Discourse Parsing
• 2019: WikiCREM: A Large Unsupervised Corpus for Coreference Resolution
• 2019: BERT for Coreference Resolution: Baselines and Analysis
• 2019: Quoref: A Reading Comprehension Dataset with Questions Requiring Coreferential Reasoning
• 2019: Coreference Resolution in Full Text Articles with BERT and Syntax-based Mention Filtering
• 2019: Cross-document coreference: An approach to capturing coreference without context
• 2020: Conundrums in Entity Coreference Resolution: Making Sense of the State of the Art Vincent Ng
• 2020: Coreferential Reasoning Learning for Language Representation
• 2020: Learning to Ignore: Long Document Coreference with Bounded Memory Neural Networks
• 2020: Revealing the Myth of Higher-Order Inference in Coreference Resolution
• 2020: Pre-training Mention Representations in Coreference Models
• 2020: Incremental Neural Coreference Resolution in Constant Memory
• 2020: CoRefi: A Crowd Sourcing Suite for Coreference Annotation
• 2020: Using Type Information to Improve Entity Coreference Resolution
• 2020: Exploring Span Representations in Neural Coreference Resolution
• 2020: Exploring Coreference Features in Heterogeneous Data
• 2020: Evaluation of Coreference Resolution Systems Under Adversarial Attacks
• 2020: Coreference for Discourse Parsing: A Neural Approach
• 2020: Adapting Coreference Resolution to Twitter Conversations
• 2020: Fast End-to-end Coreference Resolution for Korean
• 2020: Paraphrasing vs Coreferring: Two Sides of the Same Coin

pronoun

• 2017: Chinese Zero Pronoun Resolution with Deep Memory Network
• 2019: What You See is What You Get: Visual Pronoun Coreference Resolution in Dialogues
• 2020: Learning to Pronounce Chinese Without a Pronunciation Dictionary
• 2020: Does the Objective Matter? Comparing Training Objectives for Pronoun Resolution
• 2020: Transformer-GCRF: Recovering Chinese Dropped Pronouns with General Conditional Random Fields

数据收集

1. conll2012 数据
2. 千言数据集：面向中文短文本的实体链指任务 数据
3. CCKS 2021：面向通信领域的过程类知识抽取（二）事件共指消解任务 数据
4. CLUEWSC2020 数据
5. GAP-Google 数据

ontonotes数据处理

处理后： Link，或参考: Link（推荐）

开源项目整理

项目地址	框架	论文支持	备注
https://github.com/mandarjoshi90/coref	TF	2019 EMNLP	spanbert能跑通
https://github.com/kentonl/e2e-coref	TF	2017 EMNLP, 2018NAACL
https://github.com/clarkkev/deep-coref	keras	2016 EMNLP, 2016 ACL
https://github.com/ShannonAI/CorefQA	TF	2020 ACL
https://github.com/thunlp/CorefBERT	Torch	2020 EMNLP
https://github.com/google-research-datasets/gap-coreference	-	2018 ACL	GAP数据集
https://github.com/lxucs/coref-hoi	Torch	2020 ACL
https://github.com/troublemaker-r/Chinese_Coreference_Resolution	Torch	-	spanbert，作者对数据做了处理，增加了sentence_map字段，故先用自带数据测试
https://github.com/fastnlp/fastNLP/tree/master/reproduction/coreference_resolution	Torch	-	e2e

评价标准

以下内容来自于：Scoring Coreference Partitions of Predicted Mentions: A Reference Implementation

---

例：给出标准的划分和结果划分。

Label={a,b,c},{d,e},{f,g,h} 
Resolution={a,b},{c,d,e},{f,g,i}
其中abc表示一个表述，用大括号表示一个实体。处于同一个大括号的表述属于同一实体。

MUC

• Precision:指消解后的结果中，正确处理的表述占实际处理的表述的百分比
• Recall:描述消解结果中正确处理的表述占数据集中所有待消解表述的百分比

$$Precision=\frac{\sum_{i}^{N_r}(|R_i|-p(R_i))}{\sum_{i}^{N_r}(|R_i|-1)}$$ $$Recall=\frac{\sum_{i}^{N_l}(|L_i|-p(L_i))}{\sum_{i}^{N_l}(|L_i|-1)}$$

其中$L_i$表示标记数据中第i个实体的情况，$R_i$表示消解结果中第$i$个实体。‘|’表示当前实体所包含的表述数量。计算准确率的公式中的$p(R_i)$是指消解结果中第$i$个实体中实际包含了多少个真实的实体，其中没有出现在标记数据中的独立表述，我们认为它是独立的一个实体。计算召回率的公式中的$p(L_i)$是指标记数据中的第$i$个实体被划分为多少个子类。在下图中我们用实线椭圆表示标记数据的实体类，用虚线椭圆表示消解结果中的实体类。

有：

$$Precision=\frac{(2-1)+(3-2)+(3-2)}{(2-1)+(3-1)+(3-1)}=0.6$$ $$Recall=\frac{(3-2)+(2-1)+(3-2)}{(3-1)+(2-1)+(3-1)}=0.6$$

BCUB

$$Precision=\frac{\sum_{i}^{N_l} \sum_{j}^{N_r}\frac{|{L_i}\cap{R_j}|^2}{|L_i|}}{\sum_{j}^{N_r}{|L_i|}}$$ $$Recall=\frac{\sum_{i}^{N_l} \sum_{j}^{N_r}\frac{|{L_i}\cap{R_j}|^2}{|R_j|}}{\sum_{j}^{N_r}{|R_j|}}$$