信息抽取文献汇总 - peter-xbs/CommonCodes GitHub Wiki

文章汇总【持续更新】

NER

ACL2022

• CONTaiNER: Few-Shot Named Entity Recognition via Contrastive Learning

• Distantly Supervised Named Entity Recognition via Confidence-Based Multi-Class Positive and Unlabeled Learning

  • 【读后感】首先阐述了LexiconNER binary Positive and Ualabled learning任务缺点【use one-vs-all 方式，将multiclass问题转变为multi binary classification问题，有2个显著缺点：1——效率较低（特别是当entity types较多的情况下），2）？；

• Nested Named Entity Recognition with Span-level Graphs

  • sjtu，github无源码

• Packed Levitated Marker for Entity and Relation Extraction

• Parallel Instance Query Network for Named Entity Recognition

• Rethinking Negative Sampling for Handling Missing Entity Annotations

  • 【读后感】是ICLR2021 negative sampling NER的姊妹篇，negative sampling随机采样的负样本，仍有一定几率采样到unlabled postive entities；本文针对此处进行优化，在采样时优先采样label不确定或者更高概率为O的样本作为negative samples，整体上又提升了 >1个百分点；在确定weighted sampling策略，作者采用adaptive style, 迭代式的使用主训练的 NER模型来确定Po以及Puncertainty，从而来控制负样本的采样。

• Cross-domain Named Entity Recognition via Graph Matching

• Decomposed Meta-Learning for Few-Shot Named Entity Recognition

• Fusing Heterogeneous Factors with Triaffine Mechanism for Nested Named Entity Recognition

• Label Semantics for Few Shot Named Entity Recognition

• Better Modeling of Incomplete Annotations for Named Entity Recognition

NAACL2022

• Improving Entity Disambiguation by Reasoning over a Knowledge Base

• Hero-Gang Neural Model For Named Entity Recognition

• On the Use of External Data for Spoken Named Entity Recognition

• Label Refinement via Contrastive Learning for Distantly-Supervised Named Entity Recognition

• Delving Deep into Regularity: A Simple but Effective Method for Chinese Named Entity Recognition

• NER-MQMRC: Formulating Named Entity Recognition as Multi Question Machine Reading Comprehension

AAAI2022:

• W2NER: Unified Named Entity Recognition as Word-Word Relation Classification. IJCAI2022

• Position-aware Joint Entity and Relation Extraction with Attention Mechanism

• Propose-and-Refine: A Two-Stage Set Prediction Network for Nested Named Entity Recognition

• Automatic Noisy Label Correction for Fine-Grained Entity Typing COLING2022

• Simple yet Powerful: An Overlooked Architecture for Nested Named Entity Recognition

2021及以前核心论文

• Negative Sampling for NER: ICLR2021: EMPIRICAL ANALYSIS OF UNLABELED ENTITY PROBLEM IN NAMED ENTITY RECOGNITION【感觉思想很简单，但效果提升较大，同时学习到了一种新的Mask的方案，收益颇多，即计算Loss时，除了用mask想乘的方法，在sampling时，通过position循环提取所需要计算Loss的部分相加再反向传播即可】https://openreview.net/pdf?id=5jRVa89sZk

• LexiconNER ACL2019: Distantly Supervised Named Entity Recognition using Positive-Unlabeled Learning https://aclanthology.org/P19-1231.pdf; 此处PU Learning的实现还是比较令人费解的，AdaPU 的loss计算 相当于(m-p)pRisk + uRisk，有点类似于weighted Batched MAE loss，不知理解是否有误，如果是这样的话，PU Learning的意义在哪里呢？

  • UPDATE: 读了原始的

• TemplateBasedNER ACL2021: Template-Based Named Entity Recognition Using BART https://aclanthology.org/2021.findings-acl.161.pdf 此处改造比较简单，BART输入原句，decode解码生成模板，如<candidate_span> is a <entity_type> entity；后续如有需要再看细节 https://github.com/Nealcly/templateNER/blob/main/seq2seq_model.py

历史论文TODO-LIST

• Neural Chinese Named Entity Recognition via CNN-LSTM-CRF and Joint Training with Word Segmentation https://dl.acm.org/doi/abs/10.1145/3308558.3313743 www 2019

• PENNER: Pattern-enhanced Nested Named Entity Recognition in Biomedical Literature https://ieeexplore.ieee.org/abstract/document/8621485 IEEE2019

• Deep Active Learning for Named Entity Recognition

• Knowledge Guided Named Entity Recognition for BioMedical Text https://arxiv.org/abs/1911.03869 arxiv2019

• A Neural Named Entity Recognition and Multi-Type Normalization Tool for Biomedical Text Mining https://ieeexplore.ieee.org/abstract/document/8730332 IEEE2019

---

---

Relation

ACL2012

• Continual Few-shot Relation Learning via Embedding Space Regularization and Data Augmentation

• Pre-training to Match for Unified Low-shot Relation Extraction

• DoesBERTKnow that theIS-A Relation Is Transitive?

• HiCLRE: A Hierarchical Contrastive Learning Framework for Distantly Supervised Relation Extraction

• Consistent Representation Learning for Continual Relation Extraction

• Improving Relation Extraction through Syntax-induced Pre-training with Dependency Masking

• Learning from Missing Relations: Contrastive Learning with Commonsense Knowledge Graphs for Commonsense Inference

• Learning Reasoning Patterns for Relational Triple Extraction with Mutual Generation of Text and Graph

• MIMICause:Representation and automatic extraction of causal relation types from clinical notes

• Open Relation Modeling: Learning to Define Relations between Entities

• RelationPrompt: Leveraging Prompts to Generate Synthetic Data for Zero-Shot Relation Triplet Extraction

• A Simple yet Effective Relation Information Guided Approach for Few-Shot Relation Extraction

AAAI2022

• OneRel: OneRel: Joint Entity and Relation Extraction with One Module in One Step. IJCAI2022

• FastRE: Towards Fast Relation Extraction with Convolutional Encoder and Improved Cascade Binary Tagging Framework

• Relational Triple Extraction: One Step is Enough

---

---

Alignment

ACL2022

• Cross-LingualUMLSNamed Entity Linking usingUMLSDictionary Fine-Tuning

• Deep Reinforcement Learning for Entity Alignment

• Generative Biomedical Entity Linking via Knowledge Base-Guided Pre-training and Synonyms-Aware Fine-tuning https://arxiv.org/pdf/2204.05164.pdf NAACL 2022

• WikiDiverse: A Multimodal Entity Linking Dataset with Diversified Contextual Topics and Entity Types HOSMEL: A Hot-Swappable Modularized Entity Linking Toolkit for Chinese

• A Transformational Biencoder with In-Domain Negative Sampling for Zero-Shot Entity Linking

• Detection, Disambiguation, Re-ranking: Autoregressive Entity Linking as a Multi-Task Problem

• Improving Candidate Retrieval with Entity Profile Generation for Wikidata Entity Linking

• ZiNet: Linking Chinese Characters Spanning Three Thousand Years

• Graph Neural Networks for Multiparallel Word Alignment

• RuCCoN: Clinical Concept Normalization inRussian

• Generative Biomedical Entity Linking via Knowledge Base-Guided Pre-training and Synonyms-Aware Fine-tuning

• Global Entity Disambiguation with BERT

• ExtEnD: Extractive Entity Disambiguation

  • 和ESCHER比较类似，采用LongFormer解决长度过长问题

NAACL2022

• Zero-shot Entity Linking with Less Data

• Dangling-Aware Entity Alignment with Mixed High-Order Proximities

• ReFinED: An Efficient Zero-shot-capable Approach to End-to-End Entity Linking

• BLINK with Elasticsearch for Efficient Entity Linking in Business Conversations

ICLR2022

• EntQA: Entity Linking as Question Answering

AAAI2022

• Ensemble Semi-supervised Entity Alignment via Cycle-Teaching.

• DeepType 2: Superhuman Entity Linking, All You Need Is Type Interactions.

WWW2022

• Uncertainty-aware Pseudo Label Refinery for Entity Alignment

• SelfKG: Self-Supervised Entity Alignment in Knowledge Graphs

IJCAI2022

• Entity Alignment with Reliable Path Reasoning and Relation-aware Heterogeneous Graph Transformer

• Community Question Answering Entity Linking via Leveraging Auxiliary Data

• Bridging the Gap between Reality and Ideality of Entity Matching: A Revisting and Benchmark Re-Constrcution

• Enhancing Entity Representations with Prompt Learning for Biomedical Entity Linking

KDD2022

• Multi-modal Siamese Network for Entity Alignment

• ClusterEA: Scalable Entity Alignment with Stochastic Training and Normalized Mini-batch Similarities

• Few-Shot Fine-Grained Entity Typing with Automatic Label Interpretation and Instance Generation

ICML2022

• Understanding and Improving Knowledge Graph Embedding for Entity Alignment COLING2022

• Improving Zero-Shot Entity Linking Candidate Generation with Ultra-Fine Entity Type Information

历史文章收录

• GENRE ICLR2021: AUTOREGRESSIVE ENTITY RETRIEVAL https://openreview.net/pdf?id=5k8F6UU39V

  • 参考 https://zhuanlan.zhihu.com/p/365390229 生成式 受限解码

  • facebook

• BLINK ACL2020 Scalable Zero-shot Entity Linking with Dense Entity Retrieval

  • 两阶段 双塔召回+cross encoder精排

  • facebook

• ESCHER NAACL2021 ESC: RedesigningWSDwith Extractive Sense Comprehension

  • 词典语义消歧，非常简单，将input和多个候选项拼接起来，用MRC span方式选取正确候选项

EVENT

ACL2022

• A Graph EnhancedBERTModel for Event Prediction
• 一文读懂事件知识图谱

Information Extraction

ACL2022

• De-Bias for Generative Extraction in Unified NER Task
• Text-to-Table: A New Way of Information Extraction