ACL2020录取文章已经放出,链接如下:
ACL2020论文集合www.aclweb.org
为了以后更加方便地阅读论文,也本着一颗开源之心,花一个下午的时间整理了一下相关论文。鉴于本人精力有限,并且也只在NLP领域做过部分任务,因此只整理了以下四种任务的论文:(1)预训练语言模型及部分应用(18篇);(2)QA问答系统及机器阅读理解(13篇);(3)问题生成(7篇);(4)自然语言推理(文本蕴含or文本匹配)(4篇)。
欢迎大家进行补充~~~~~~
一、预训练语言模型及部分应用
(1)QuASE: Question-Answer Driven Sentence Encoding
paper:论文; code:github
(2)TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data
paper:论文; code:github
(3)Don’t Stop Pretraining: Adapt Language Models to Domains and Tasks
paper:论文; code:github
(4)BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension
paper:论文; code:github
(5)Toward Better Storylines with Sentence-Level Language Models
paper:论文; code:github
(6)tBERT: Topic Models and BERT Joining Forces for Semantic Similarity Detection
paper:论文; code:github
(7)FastBERT: a Self-distilling BERT with Adaptive Inference Time
paper:论文; code:github
(8)Pretraining with Contrastive Sentence Objectives Improves Discourse Performance of Language Models
paper:论文; code:github
(9)DeFormer: Decomposing Pre-trained Transformers for Faster Question Answering
paper:论文; code:github
(10)Enhancing Pre-trained Chinese Character Representation with Word-aligned Attention
paper:论文; code:github
(11)schuBERT: Optimizing Elements of BERT
paper:论文
(12)Span Selection Pre-training for Question Answering
paper:论文; code:github
(13)DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference
paper:论文; code:github
(14)MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices
paper:论文; code:github
(15)A Generative Model for Joint Natural Language Understanding and Generation
paper:论文
(16)Fast and Accurate Deep Bidirectional Language Representations for Unsupervised Learning
paper:论文; code:github
(17)Probabilistically Masked Language Model Capable of Autoregressive Generation in Arbitrary Word Order
paper:论文
(18)Few-Shot NLG with Pre-Trained Language Model
paper:论文; code:github
二、QA问答系统及机器阅读理解
(1)DoQA - Accessing Domain-Specific FAQs via Conversational QA
paper:论文; data:网页
(2)CorefQA: Coreference Resolution as Query-based Span Prediction
paper:论文; code:github
(3)RikiNet: Reading Wikipedia Pages for Natural Question Answering
paper:论文
(4)Harvesting and Refining Question-Answer Pairs for Unsupervised QA
paper:论文; code:github
(5)Probabilistic Assumptions Matter: Improved Models for Distantly-Supervised Document-Level Question Answering
paper:论文; code:github
(6)Template-Based Question Generation from Retrieved Sentences for Improved Unsupervised Question Answering
paper:论文; code:github
(7)Contextualized Sparse Representations for Real-Time Open-Domain Question Answering
paper:论文; code:github
(8)Recurrent Chunking Mechanisms for Long-Text Machine Reading Comprehension
paper:论文; code:github
(9)Document Modeling with Graph Attention Networks for Multi-grained Machine Reading Comprehension
paper:论文; code:github
(10)Unsupervised Alignment-based Iterative Evidence Retrieval for Multi-hop Question Answering
paper:论文; code:github
(11)Improving Multi-hop Question Answering over Knowledge Graphs using Knowledge Base Embeddings
paper:论文; code:github
(12)A Self-Training Method for Machine Reading Comprehension with Soft Evidence Extraction
paper:论文
(13)A Frame-based Sentence Representation for Machine Reading Comprehension
paper:论文
三、问题生成
(1)ClarQ: A large-scale and diverse dataset for Clarification Question Generation
paper:论文; code:github
(2)How to Ask Good Questions? Try to Leverage Paraphrases
paper:论文
(3)On the Importance of Diversity in Question Generation for QA
paper:论文
(4)Semantic Graphs for Generating Deep Questions
paper:论文; code:github
(5)Unsupervised FAQ Retrieval with Question Generation and BERT
paper:论文; code:github
(6)Syn-QG: Syntactic and Shallow Semantic Rules for Question Generation
paper:论文; code:github
(7)Low-Resource Generation of Multi-hop Reasoning Questions
paper:论文
四、自然语言推理(文本蕴含or文本匹配)
(1)Uncertain Natural Language Inference
paper:论文; code:github
(2)Towards Robustifying NLI Models Against Lexical Dataset Biases
paper:论文; code:github
(3)NILE : Natural Language Inference with Faithful Natural Language Explanations
paper:论文; code:github
(4)Rationalizing Text Matching: Learning Sparse Alignments via Optimal Transport
paper:论文; code:github
ACL2020共收录了779篇论文,整理下来实属不易,还请大家多多点赞~~~
鉴于本人仅对上述几个任务有所涉及,因此只整理了这几个模块;其他小伙伴如果有一些建议或愿意一起整理,可以留言,会持续更新~~~~
接下来,有可能会进一步进行解读,将论文摘要部分添加进来,给更多小伙伴提供方便。
广告时间,哈哈哈哈。推荐几篇本人之前写的一些文章: