MeteorMan将在本文针对ACL2020开源论文分专题介绍。
如今ACL2020已结束，各大论文已经放出，眼热的我针对里面的论文资源依据个人兴趣分门别类整理，并特别针对开源论文进一步处理，将其划分为如下4大专题：
- 问答系统和阅读理解
- 问题生成
- 自然语言推理
- 预训练语言模型及应用
1.QA问答系统及机器阅读理解
Harvesting and Refining Question-Answer Pairs for Unsupervised QA
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.600.pdf
- 代码链接：https://github.com/Neutralzz/RefQA
Probabilistic Assumptions Matter: Improved Models for Distantly-Supervised Document-Level Question Answering
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.501.pdf
- 代码链接：https://github.com/hao-cheng/ds_doc_qa
Template-Based Question Generation from Retrieved Sentences for Improved Unsupervised Question Answering
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.413.pdf
- 代码链接：https://github.com/awslabs/unsupervised-qa
Contextualized Sparse Representations for Real-Time Open-Domain Question Answering
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.85.pdf
- 代码链接：https://github.com/jhyuklee/sparc
Recurrent Chunking Mechanisms for Long-Text Machine Reading Comprehension
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.603.pdf
- 代码链接：https://github.com/HongyuGong/RCM-Question-Answering.git
Document Modeling with Graph Attention Networks for Multi-grained Machine Reading Comprehension
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.599.pdf
- 代码链接：https://github.com/DancingSoul/NQ_BERT-DM
Unsupervised Alignment-based Iterative Evidence Retrieval for Multi-hop Question Answering
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.414.pdf
- 代码链接：https://github.com/vikas95/AIR-retriever
Improving Multi-hop Question Answering over Knowledge Graphs using Knowledge Base Embeddings
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.412.pdf
- 代码链接：https://github.com/malllabiisc/EmbedKGQA
2.问题生成
ClarQ: A large-scale and diverse dataset for Clarification Question Generation
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.651.pdf
- 代码链接：https://github.com/vaibhav4595/ClarQ
Semantic Graphs for Generating Deep Questions
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.135.pdf
- 代码链接：https://github.com/WING-NUS/SG-Deep-Question-Generation
Syn-QG: Syntactic and Shallow Semantic Rules for Question Generation
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.69.pdf
- 代码链接：https://bitbucket.org/kaustubhdhole/syn-qg/
Unsupervised FAQ Retrieval with Question Generation and BERT
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.74.pdf
- 代码链接：https://github.com/YosiMass/faq-retrieval
3.自然语言推理（文本蕴含or文本匹配）
Towards Robustifying NLI Models Against Lexical Dataset Biases
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.773.pdf
- 代码链接：https://github.com/owenzx/LexicalDebias-ACL2020
NILE : Natural Language Inference with Faithful Natural Language Explanations
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.771.pdf
- 代码链接：https://github.com/SawanKumar28/nile
Extracting Headless MWEs from Dependency Parse Trees: Parsing, Tagging, and Joint Modeling Approaches
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.775.pdf
- 代码链接：https://github.com/tzshi/flat-mwe-parsing
Uncertain Natural Language Inference
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.774.pdf
- 代码链接：https://github.com/ctongfei/unli
4.预训练语言模型及部分应用
QuASE: Question-Answer Driven Sentence Encoding
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.772.pdf
- 代码链接：https://github.com/CogComp/QuASE
TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.745.pdf
- 代码链接：https://github.com/facebookresearch/TaBERT
Don’t Stop Pretraining: Adapt Language Models to Domains and Tasks
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.740.pdf
- 代码链接：https://github.com/allenai/dont-stop-pretraining
BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.703.pdf
- 代码链接：https://github.com/pytorch/fairseq
Toward Better Storylines with Sentence-Level Language Models
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.666.pdf
- 代码链接：https://github.com/google-research/google-research/tree/master/better_storylines
tBERT: Topic Models and BERT Joining Forces for Semantic Similarity Detection
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.630.pdf
- 代码链接：https://github.com/wuningxi/tBERT
FastBERT: a Self-distilling BERT with Adaptive Inference Time
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.537.pdf
- 代码链接：https://github.com/autoliuweijie/FastBERT
Pretraining with Contrastive Sentence Objectives Improves Discourse Performance of Language Models
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.439.pdf
- 代码链接：https://github.com/google-research/language/tree/master/language/conpono
DeFormer: Decomposing Pre-trained Transformers for Faster Question Answering
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.411.pdf
- 代码链接：https://github.com/StonyBrookNLP/deformer
Enhancing Pre-trained Chinese Character Representation with Word-aligned Attention
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.315.pdf
- 代码链接：https://github.com/lsvih/MWA
Span Selection Pre-training for Question Answering
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.247.pdf
- 代码链接：https://github.com/IBM/span-selection-pretraining
DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.204.pdf
- 代码链接：https://github.com/castorini/DeeBERT
MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.195.pdf
- 代码链接：https://github.com/google-research/google-research/tree/master/mobilebert
Fast and Accurate Deep Bidirectional Language Representations for Unsupervised Learning
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.76.pdf
- 代码链接：https://github.com/joongbo/tta
Few-Shot NLG with Pre-Trained Language Model
- 论文链接：https://www.aclweb.org/anthology/2020.acl-main.18.pdf
- 代码链接：https://github.com/czyssrs/Few-Shot-NLG
