├── LICENSE ├── README.md ├── kmrc-area.md ├── v0.md └── v1.md /LICENSE: -------------------------------------------------------------------------------- 1 | MIT License 2 | 3 | Copyright (c) 2018 Yuqiang Xie 4 | 5 | Permission is hereby granted, free of charge, to any person obtaining a copy 6 | of this software and associated documentation files (the "Software"), to deal 7 | in the Software without restriction, including without limitation the rights 8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 9 | copies of the Software, and to permit persons to whom the Software is 10 | furnished to do so, subject to the following conditions: 11 | 12 | The above copyright notice and this permission notice shall be included in all 13 | copies or substantial portions of the Software. 14 | 15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 21 | SOFTWARE. 22 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Papers on Knowledge-based Machine Reading Comprehension. 2 | 3 | A list of recent papers about **Knowledge-based Machine Reading Comprehension** (**KMRC**). 4 | 5 | Contributed by [Luxi Xing](https://github.com/XingLuxi) and [Yuqiang Xie](https://github.com/IndexFziQ). 6 | 7 | Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China. 8 | 9 | Update on **Oct. 23, 2019**. 10 | 11 | (the current version only contains the works published on the conferences or journals, we will continuously update this list.) 12 | 13 | ***Because of some mistakes, this version can not be updated. For the latest version, please jump to [KMRC-Research-Archive](https://github.com/XingLuxi/KMRC-Research-Archive).*** 14 | 15 | ------- 16 | 17 | ## [Content](#content) 18 | 19 | 1. [Survey](#survey-papers) 20 | 2. [Cloze Style Tasks](#cloze-style-tasks) 21 | 3. [Span Extraction Tasks](#span-extraction-tasks) 22 | 4. [Multiple Choice Tasks](#multiple-choice-tasks) 23 | 5. [Generation Tasks](#generation-tasks) 24 | 6. [Datasets](#benchmark-datasets) 25 | 26 | 27 | ## [Survey Papers](#content) 28 | 29 | 1. **Commonsense Reasoning for Natural Language Understanding: A Survey of Benchmarks, Resources, and Approaches**. 2019. [[paper](https://arxiv.org/abs/1904.01172) / [note](https://zhuanlan.zhihu.com/p/80883568)] 30 | 31 | Authors: *Shane Storks, Qianzi Gao, Joyce Y. Chai* 32 | 33 | 1. **Neural Machine Reading Comprehension: Methods and Trends**. 2019. [[paper](https://arxiv.org/abs/1907.01118)] 34 | 35 | Authors: *Shanshan Liu, Xin Zhang, Sheng Zhang, Hui Wang, Weiming Zhang* 36 | 37 | 1. **Machine Reading Comprehension: a Literature Review**. 2019. [[paper](https://arxiv.org/abs/1907.01686)] 38 | 39 | Authors: *Xin Zhang, An Yang, Sujian Li, Yizhong Wang* 40 | 41 | 42 | ## [Cloze Style Tasks](#content) 43 | 44 | | Title | Publish | Tasks | Links | 45 | | :---- | :---: | :---- | :---:| 46 | | Reasoning with Heterogeneous Knowledge for Commonsense Machine Comprehension | ACL
2017 | SCT | [paper](http://aclweb.org/anthology/D17-1216) | 47 | | World Knowledge for Reading Comprehension: Rare Entity Prediction with Hierarchical LSTMs Using External Descriptions | EMNLP
2017 | Rare Entity Prediction | [paper](https://www.aclweb.org/anthology/D17-1086) | 48 | | Knowledgeable Reader: Enhancing Cloze-Style Reading Comprehension with External Commonsense Knowledge | ACL
2018 | Common Nouns | [paper](http://aclweb.org/anthology/P18-1076)
[note](http://xingluxi.github.io/2019/01/09/paper-knreader/) | 49 | | A Multi-Attention based Neural Network with External Knowledge for Story Ending Predicting Task | COLING
2018 | SCT | [paper](http://www.aclweb.org/anthology/C18-1149) | 50 | | Incorporating Structured Commonsense Knowledge in Story Completion | AAAI
2018 | SCT | [paper](https://arxiv.org/abs/1811.00625) | 51 | | Story Ending Prediction by Transferable BERT | IJCAI
2019 | SCT | [paper](https://arxiv.org/abs/1905.07504) | 52 | 53 | 54 | 55 | ## [Span Extraction Tasks](#content) 56 | 57 | | Title | Publish | Tasks | Links | 58 | | :---- | :---: | :---- | :---:| 59 | | Dynamic Integration of Background Knowledge in Neural NLU Systems | 2018 | SQuAD/
TriviaQA | [paper](https://arxiv.org/pdf/1706.02596.pdf)
[note](http://xingluxi.github.io/2019/03/06/paper-2018-refinewordemb/) | 60 | | Explicit Utilization of General Knowledge in Machine Reading Comprehension | ACL
2019 | SQuAD | [paper](https://www.aclweb.org/anthology/P19-1219)
[note](http://xingluxi.github.io/2019/07/17/paper-acl2019-kar/) | 61 | | Enhancing Pre-Trained Language Representations with Rich Knowledge for Machine Reading Comprehension | ACL
2019 | SQuAD/
ReCoRD | [paper](https://www.aclweb.org/anthology/P19-1226/)
[note](http://xingluxi.github.io/2019/07/29/paper-acl2019-kt-net/) | 62 | 63 | 64 | ## [Multiple Choice Tasks](#content) 65 | 66 | | Title | Publish | Tasks | Links | 67 | | :---- | :---: | :---- | :---:| 68 | | Yuanfudao at SemEval-2018 Task 11: Three-way Attention and Relational Knowledge for Commonsense Machine Comprehension | SemEval
2018 | SemEval-2018 Task 11 | [paper](https://www.aclweb.org/anthology/S18-1120/)
[code](https://github.com/intfloat/commonsense-rc) | 69 | | Improving Question Answering by Commonsense-Based Pre-Training | AAAI
2019 | ARC/
OpenBookQA/
SemEval-2018 Task 11 | [paper](https://arxiv.org/abs/1809.03568) | 70 | | Improving Machine Reading Comprehension with General Reading Strategies | NAACL
2019 | ARC/ OpenBookQA/ MCTest/
SemEval-2018 Task 11/ SCT/ MultiRC | [paper](https://www.aclweb.org/anthology/N19-1270/)
[code](https://github.com/nlpdata/strategy/) | 71 | | Ranking and Selecting Multi-Hop Knowledge Paths to Better Predict Human Needs | NAACL
2019 | story commonsense | [paper](https://www.aclweb.org/anthology/papers/N/N19/N19-1368/)
[code](https://github.com/debjitpaul/Multi-Hop-Knowledge-Paths-Human-Needs) | 72 | | Incorporating Relation Knowledge into Commonsense Reading Comprehension with Multi-task Learning | CIKM
2019 | SemEval-2018 Task 11 / SCT | [paper](https://arxiv.org/abs/1908.04530) | 73 | | Explain Yourself! Leveraging Language Models for Commonsense Reasoning | ACL
2019 | CommonsenseQA | [paper](https://www.aclweb.org/anthology/P19-1487.pdf)
[code](https://github.com/salesforce/cos-e)
[note](http://xingluxi.github.io/2019/07/10/paper-acl2019-cos-e/) | 74 | | Careful Selection of Knowledge to solve Open Book Question Answering | ACL
2019 | OpenBookQA | [paper](https://arxiv.org/abs/1907.10738) | 75 | | Improving Question Answering with External Knowledge | EMNLP
MRQA
2019 | ARC/
OpenBookQA | [paper](https://arxiv.org/pdf/1902.00993.pdf)
[note](http://xingluxi.github.io/2019/08/26/paper-2019-edl-md/) | 76 | | KagNet: Knowledge-Aware Graph Networks for Commonsense Reasoning | EMNLP
2019 | CommonsenseQA | [paper](https://arxiv.org/abs/1909.02151)
[code](https://github.com/INK-USC/KagNet)
[note](https://zhuanlan.zhihu.com/p/81917730) | 77 | | What’s Missing: A Knowledge Gap Guided Approach for Multi-hop Question Answering | EMNLP
2019 | OpenBookQA | [paper](https://arxiv.org/abs/1909.09253)
[note](https://zhuanlan.zhihu.com/p/85852386) | 78 | | BIG MOOD: Relating Transformers to Explicit Commonsense Knowledge | EMNLP
COIN
2019 | MCScripts v2 | [paper](http://arxiv.org/abs/1910.07713) | 79 | 80 | 81 | 82 | ## [Generation Tasks](#content) 83 | Also known as **Free-form Answer Tasks** 84 | 85 | | Title | Publish | Tasks | Links | 86 | | :---- | :---: | :---- | :---:| 87 | | Commonsense for Generative Multi-Hop Question Answering Tasks | EMNLP
2018 | NarrativeQA/
WikiHop | [paper](https://www.aclweb.org/anthology/D18-1454/)
[code](https://github.com/yicheng-w/CommonSenseMultiHopQA)
[note](http://xingluxi.github.io/2019/02/21/paper-emnlp2018-mhpgm/) | 88 | | COMET: Commonsense Transformers for Automatic Knowledge Graph Construction | ACL
2019 | Atomic | [paper](https://arxiv.org/abs/1906.05317)
[code](https://github.com/atcbosselut/comet-commonsense)
[note](https://indexfziq.github.io/2019/07/03/COMET/) | 89 | | Incorporating External Knowledge into Machine Reading for Generative Question Answering | EMNLP
2019 | MS MARCO | [paper](http://arxiv.org/abs/1909.02745)
[note](http://xingluxi.github.io/2019/10/13/paper-emnlp2019-keag/) | 90 | 91 | 92 | ## [Benchmark Datasets](#content) 93 | 94 | 1. [**COPA**] Choice of Plausible Alternatives: An Evaluation of Commonsense Causal Reasoning. AAAI,2011. [[paper](http://ict.usc.edu/~gordon/publications/AAAI-SPRING11A.PDF) / [data](http://people.ict.usc.edu/~gordon/copa.html)] 95 | 96 | Authors: *Melissa Roemmele, Cosmin Adrian Bejan, Andrew S. Gordon* 97 | 98 | * Type: Multiple-Choice; 99 | 100 | 1. [**WSC**] The Winograd Schema Challenge. AAAI,2011. [[paper](https://www.aaai.org/ocs/index.php/KR/KR12/paper/download/4492/4924) /[data](https://cs.nyu.edu/faculty/davise/papers/WinogradSchemas/WS.html)] 101 | 102 | Authors: *Hector J. Levesque, Ernest Davis, Leora Morgenstern* 103 | 104 | * Type: Multiple-Choice; 105 | 106 | 2. [**ROCStories**; **SCT**] A Corpus and Cloze Evaluation for Deeper Understanding of Commonsense Stories. NAACL,2016. [[paper](https://www.aclweb.org/anthology/N16-1098/) / [data](https://www.cs.rochester.edu/nlp/rocstories/)] 107 | 108 | Authors: *Nasrin Mostafazadeh, Nathanael Chambers, Xiaodong He, Devi Parikh, Dhruv Batra, Lucy Vanderwende, Pushmeet Kohli, James Allen* 109 | 110 | * Type: Cloze; 111 | 112 | 1. [**NarrativeQA**] The NarrativeQA Reading Comprehension Challenge. TACL,2018. [[paper](https://arxiv.org/abs/1712.07040) / [data](https://github.com/deepmind/narrativeqa)] 113 | 114 | Authors: *Tomáš Kočiský, Jonathan Schwarz, Phil Blunsom, Chris Dyer, Karl Moritz Hermann, Gábor Melis, Edward Grefenstette* 115 | 116 | * Type: Generation; 117 | 118 | 1. [**SemEval-2018 Task 11**] MCScript: A Novel Dataset for Assessing Machine Comprehension Using Script Knowledge. LERC,2018. [[paper](https://arxiv.org/abs/1803.05223) / [data](http://www.sfb1102.uni-saarland.de/?page_id=2582)] 119 | 120 | Authors: *Simon Ostermann, Ashutosh Modi, Michael Roth, Stefan Thater, Manfred Pinkal* 121 | 122 | * Type: Multiple-Choice; 123 | 124 | 1. [**story-commonsense**] Modeling Naive Psychology of Characters in Simple Commonsense Stories. ACL,2018. [[paper](https://www.aclweb.org/anthology/P18-1213/) / [data](http://uwnlp.github.io/storycommonsense)] 125 | 126 | Authors: *Hannah Rashkin, Antoine Bosselut, Maarten Sap, Kevin Knight, Yejin Choi* 127 | 128 | * Type: Multiple-Choice; 129 | 130 | 1. **Event2Mind**: Commonsense Inference on Events, Intents, and Reactions. ACL,2018. [[paper](https://www.aclweb.org/anthology/P18-1043/) / [data](https://uwnlp.github.io/event2mind/)] 131 | 132 | Authors: *Hannah Rashkin, Maarten Sap, Emily Allaway, Noah A. Smith, Yejin Choi* 133 | 134 | * Types: Generation; 135 | 136 | 1. **ATOMIC**: An Atlas of Machine Commonsense for If-Then Reasoning. AAAI,2019. [[paper](https://homes.cs.washington.edu/~msap/atomic/data/sap2019atomic.pdf) / [data](https://homes.cs.washington.edu/~msap/atomic/)] 137 | 138 | Authors: *Maarten Sap, Ronan LeBras, Emily Allaway, Chandra Bhagavatula, Nicholas Lourie, Hannah Rashkin, Brendan Roof, Noah A. Smith, Yejin Choi* 139 | 140 | * Types: Generation; 141 | 142 | 1. [**ARC**] Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge. 2018. [[paper](http://ai2-website.s3.amazonaws.com/publications/AI2ReasoningChallenge2018.pdf) / [data](http://data.allenai.org/arc/)] 143 | 144 | Authors: *Peter Clark, Isaac Cowhey, Oren Etzioni, Tushar Khot, Ashish Sabharwal, Carissa Schoenick, Oyvind Tafjord* 145 | 146 | * Type: Multiple-Choice; 147 | 148 | 1. [**OpenBookQA**] Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answering. EMNLP,2018. [[paper](https://www.aclweb.org/anthology/D18-1260/) / [data](https://leaderboard.allenai.org/open_book_qa)] 149 | 150 | Authors: *Todor Mihaylov, Peter Clark, Tushar Khot, Ashish Sabharwal* 151 | 152 | * Type: Multiple-Choice; 153 | 154 | 1. **ReCoRD**: Bridging the Gap between Human and Machine Commonsense Reading Comprehension. 2018. [[paper](https://arxiv.org/abs/1810.12885) / [data](https://sheng-z.github.io/ReCoRD-explorer/)] 155 | 156 | Authors: *Sheng Zhang, Xiaodong Liu, Jingjing Liu, Jianfeng Gao, Kevin Duh, Benjamin Van Durme* 157 | 158 | * Type: Cloze; 159 | 160 | 1. **CommonsenseQA**: A Question Answering Challenge Targeting Commonsense Knowledge. NAACL,2019. [[paper](https://www.aclweb.org/anthology/N19-1421/) / [data](https://www.tau-nlp.org/commonsenseqa)] 161 | 162 | Authors: *Alon Talmor, Jonathan Herzig, Nicholas Lourie, Jonathan Berant* 163 | 164 | * Type: Multiple-Choice; 165 | 166 | 1. **ChID**: A Large-scale Chinese IDiom Dataset for Cloze Test. ACL,2019. [[paper](https://www.aclweb.org/anthology/P19-1075) / [data](https://github.com/chujiezheng/ChID-Dataset)] 167 | 168 | Authors: *Chujie Zheng, Minlie Huang, Aixin Sun* 169 | 170 | * Type: Cloze; 171 | 172 | 1. [**sense-making**] Does it Make Sense? And Why? A Pilot Study for Sense Making and Explanation. ACL,2019. [[paper](https://www.aclweb.org/anthology/P19-1393/) / [data]( https://github.com/wangcunxiang/SenMaking-and-Explanation)] 173 | 174 | Authors: *Cunxiang Wang, Shuailong Liang, Yue Zhang, Xiaonan Li, Tian Gao* 175 | 176 | * Type: Multiple-Choice; 177 | 178 | 1. **HellaSwag**: Can a Machine Really Finish Your Sentence? ACL,2019. [[paper](https://arxiv.org/abs/1905.07830) / [data](https://rowanzellers.com/hellaswag/)] 179 | 180 | Authors: *Rowan Zellers, Ari Holtzman, Yonatan Bisk, Ali Farhadi, Yejin Choi* 181 | 182 | * Type: Multiple-Choice; 183 | 184 | 1. **SocialIQA**: Commonsense Reasoning about Social Interactions. EMNLP,2019. [[paper](https://arxiv.org/abs/1904.09728) / [data](https://maartensap.github.io/social-iqa/)] 185 | 186 | Authors: *Maarten Sap, Hannah Rashkin, Derek Chen, Ronan LeBras, Yejin Choi* 187 | 188 | * Type: Multiple-Choice; 189 | 190 | 1. [**ANLI**] Abductive Commonsense Reasoning. 2019. [[paper](https://arxiv.org/abs/1908.05739) / [data](https://leaderboard.allenai.org/anli/submissions/get-started)] 191 | 192 | Authors: *Chandra Bhagavatula, Ronan Le Bras, Chaitanya Malaviya, Keisuke Sakaguchi, Ari Holtzman, Hannah Rashkin, Doug Downey, Scott Wen-tau Yih, Yejin Choi* 193 | 194 | * Type: Multiple-Choice; 195 | 196 | 1. **Cosmos QA**: Machine Reading Comprehension with Contextual Commonsense Reasoning. EMNLP,2019. [[paper](https://arxiv.org/abs/1909.00277) / [data](https://wilburone.github.io/cosmos/)] 197 | 198 | Authors: *Lifu Huang, Ronan Le Bras, Chandra Bhagavatula, Yejin Choi* 199 | 200 | * Type: Multiple-Choice; 201 | 202 | Note: *Only consider the benchmark datasets/tasks which require knowledge to complete.* 203 | 204 | 205 | ## Other Paper List About MRC 206 | 207 | [thunlp/RCPapers](https://github.com/thunlp/RCPapers) 208 | -------------------------------------------------------------------------------- /kmrc-area.md: -------------------------------------------------------------------------------- 1 | # MRC with Knowledge 2 | 3 | | Conf. | Title | Authors/Org. | Note | 4 | | :-----: | :--- | :----------- | :---:| 5 | | ACL
2017 | [Leveraging knowledge bases in lstms for improving machine reading](https://doi.org/10.18653/v1/P17-1132) | Yang, et al.
CMU | | 6 | | ACL
2017 | [Reasoning with Heterogeneous Knowledge for Commonsense Machine Comprehension](https://www.aclweb.org/anthology/D17-1216) | Hongyu Lin, et al. | | 7 | | ACL
2017 | [World knowledge for reading comprehension: Rare entity prediction with hierarchical lstms using external descriptions](http://www.aclweb.org/anthology/D17-1086) | Long, et al.
McGill University| | 8 | | ACL
2018 | [Knowledgeable Reader: Enhancing Cloze-Style Reading Comprehension with External Commonsense Knowledge](http://aclweb.org/anthology/P18-1076) | Mihaylov, et al.
Heidelberg University | | 9 | | 2018 | [Dynamic Integration of Background Knowledge in Neural NLU Systems](https://arxiv.org/pdf/1706.02596.pdf) | Dirk Weissenborn, et al. | [blog](http://xingluxi.github.io/2019/03/06/paper-2018-refinewordemb/) | 10 | | EMNLP
2018 | [Commonsense for Generative Multi-Hop Question Answering Tasks](http://aclweb.org/anthology/D18-1454) | Lisa Bauer | [blog](http://xingluxi.github.io/2019/02/21/paper-emnlp2018-mhpgm/) | 11 | | EMNLP
2018 | [Open Domain Question Answering Using Early Fusion of Knowledge Bases and Text](https://aclweb.org/anthology/D18-1455) | CMU | | 12 | | 2018 | [Exploring Machine Reading Comprehension with Explicit Knowledge](https://arxiv.org/abs/1809.03449) | Emory University | | 13 | | ACL
2018 | [Open Domain Question Answering Using Early Fusion of Knowledge Bases and Text](https://aclweb.org/anthology/D18-1455) | CMU | | 14 | | AAAI
2018 | [Incorporating Structured Commonsense Knowledge in Story Completion](https://arxiv.org/abs/1811.00625) | | | 15 | | 2019 | [Improving Question Answering with External Knowledge](https://arxiv.org/abs/1902.00993) | | | 16 | | | | | | 17 | 18 | 19 | ## NLI with Knowledge 20 | 21 | | Conf. | Title | Authors/Org. | Note | 22 | | :-----: | :--- | :----------- | :---:| 23 | | ACL
2018 | [Neural Natural Language Inference Models Enhanced with External Knowledge](http://www.aclweb.org/anthology/P18-1224) | Qian Chen | [link](https://github.com/XingLuxi/KMRC-Papers/blob/master/note/kim.md) | 24 | 25 | ## MRC with Knowledge 26 | 27 | | Conf. | Title | Authors/Org. | Note | 28 | | :-----: | :--- | :----------- | :---:| 29 | | ACL
2017 | [Leveraging knowledge bases in lstms for improving machine reading](https://doi.org/10.18653/v1/P17-1132) | Yang, et al.
CMU | | 30 | | ACL
2017 | [World knowledge for reading comprehension: Rare entity prediction with hierarchical lstms using external descriptions](http://www.aclweb.org/anthology/D17-1086) | Long, et al.
McGill University| | 31 | | ACL
2018 | [Knowledgeable Reader: Enhancing Cloze-Style Reading Comprehension with External Commonsense Knowledge](http://aclweb.org/anthology/P18-1076) | Mihaylov, et al.
Heidelberg University | [link](https://github.com/XingLuxi/KMRC-Papers/blob/master/note/knreader.md) | 32 | | AAAI
2019 | [Story Ending Generation with Incremental Encoding and Commonsense Knowledge](https://arxiv.org/abs/1808.10113) | Jian Guan, et al.
THU | | 33 | 34 | More details refers to this [link](https://github.com/XingLuxi/KMRC-Papers/blob/master/kmrc-area.md) 35 | 36 | ## Dialog with Knowledge 37 | 38 | | Conf. | Title | Authors/Org. | Note | 39 | | :-----: | :--- | :----------- | :---:| 40 | | ACL
2017 | [Learning Symmetric Collaborative Dialogue Agents with Dynamic Knowledge Graph Embeddings](http://aclweb.org/anthology/P17-1162) | He, et al.
Stanford | | 41 | | IJCAI18 | [Commonsense Knowledge Aware Conversation Generation with Graph Attention](https://www.ijcai.org/proceedings/2018/0643.pdf) | Zhou, et al.
THU | | 42 | | AAAI
2018 | [A Knowledge-Grounded Neural Conversation Model](https://www.microsoft.com/en-us/research/wp-content/uploads/2017/02/A_Knowledge_Grounded_Neural_Conversation_Model.pdf) | Ghazvininejad, et al.
Microsoft Research | | 43 | | AAAI
2018 | [Flexible End-to-End Dialogue System for Knowledge Grounded Conversation](https://arxiv.org/pdf/1709.04264.pdf) | Zhu, et al.
HKUST | | 44 | | 2019 | Learning to Select Knowledge for Response Generation in Dialog Systems | Baidu | | 45 | 46 | 47 | 48 | ## Representation with Knowledge 49 | 50 | | Conf. | Title | Authors/Org. | Note | 51 | | :-----: | :--- | :----------- | :---:| 52 | | ICLR
2017 | [A Neural Knowledge Language Model](https://arxiv.org/pdf/1608.00318v1.pdf) | Ahn.et al.
Université de Montréal | | 53 | | ACL
2017 | [Improved Word Representation Learning with Sememes](http://aclweb.org/anthology/P17-1187) | Niu, et al.
THU | | 54 | | ACL
2018 | [Cross-lingual Lexical Sememe Prediction](http://aclweb.org/anthology/D18-1033) | Qi, et al.
THU| | 55 | | AAAI
2018 | [Improving Neural Fine-Grained Entity Typing with Knowledge Attention](https://aaai.org/ocs/index.php/AAAI/AAAI18/paper/view/16321/16167) | Xin, et al.
THU | | 56 | 57 | 58 | #### Level: 59 | 1. KMRC area. 60 | 2. Relevant research area. 61 | 3. Heuristic. 62 | 4. Review. 63 | 64 | #### Field: 65 | - **KMRC:** **K**nowledge-based **M**achine **R**eading **C**omprehension; 66 | - **KDS:** **K**nowledge-based **D**ialogue **S**ystem; 67 | - **KIR:** **K**nowledge-based **I**nformation **R**etrieval; 68 | - **SC:** **S**ememe **C**omputation; 69 | - **NKLM:** **N**eural **K**nowledge **L**anguage **M**odel. 70 | 71 | -------------------------------------------------------------------------------- /v0.md: -------------------------------------------------------------------------------- 1 | # Papers on Knowledge-based Machine Reading Comprehension. 2 | 3 | A list of recent papers about **Knowledge-based Machine Reading Comprehension** (**KMRC**). 4 | 5 | Contributed by [Luxi Xing](https://github.com/XingLuxi) and [Yuqiang Xie](https://github.com/IndexFziQ). 6 | 7 | Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China. 8 | 9 | Update on **Oct. 23, 2019**. 10 | 11 | (the current version only contains the works published on the conferences or journals, we will continuously update this list.) 12 | 13 | ------- 14 | 15 | ## [Content](#content) 16 | 17 | 1. [Survey](#survey-papers) 18 | 2. [Datasets](#benchmark-datasets) 19 | 3. [Cloze Style Tasks](#cloze-style-tasks) 20 | 4. [Span Extraction Tasks](#span-extraction-tasks) 21 | 5. [Multiple Choice Tasks](#multiple-choice-tasks) 22 | 6. [Generation Tasks](#generation-tasks) 23 | 24 | Note: papers about KBQA will be not included in this list. 25 | 26 | ## [Survey Papers](#content) 27 | 28 | 1. **Commonsense Reasoning for Natural Language Understanding: A Survey of Benchmarks, Resources, and Approaches**. 2019. 29 | 30 | Authors: *Shane Storks, Qianzi Gao, Joyce Y. Chai* 31 | 32 | Links: [paper](https://arxiv.org/abs/1904.01172) / [note](https://zhuanlan.zhihu.com/p/80883568) 33 | 34 | 1. **Neural Machine Reading Comprehension: Methods and Trends**. 2019. 35 | 36 | Authors: *Shanshan Liu, Xin Zhang, Sheng Zhang, Hui Wang, Weiming Zhang* 37 | 38 | 1. **Machine Reading Comprehension: a Literature Review**. 2019. 39 | 40 | Authors: *Xin Zhang, An Yang, Sujian Li, Yizhong Wang* 41 | 42 | Links: [paper](https://arxiv.org/abs/1907.01686) 43 | 44 | ## [Benchmark Datasets](#content) 45 | 46 | 1. [**COPA**] Choice of Plausible Alternatives: An Evaluation of Commonsense Causal Reasoning. AAAI,2011. 47 | 48 | Authors: *Melissa Roemmele, Cosmin Adrian Bejan, Andrew S. Gordon* 49 | 50 | Links: [paper](http://ict.usc.edu/~gordon/publications/AAAI-SPRING11A.PDF) / [data](http://people.ict.usc.edu/~gordon/copa.html) 51 | 52 | Type: Multiple-Choice; 53 | 54 | 1. [**WSC**] The Winograd Schema Challenge. AAAI,2011 55 | 56 | Authors: *Hector J. Levesque, Ernest Davis, Leora Morgenstern* 57 | 58 | Links: [paper](https://www.aaai.org/ocs/index.php/KR/KR12/paper/download/4492/4924) /[data](https://cs.nyu.edu/faculty/davise/papers/WinogradSchemas/WS.html) 59 | 60 | Type: Multiple-Choice; 61 | 62 | 2. [**ROCStories**; **SCT**] A Corpus and Cloze Evaluation for Deeper Understanding of Commonsense Stories. NAACL,2016. 63 | 64 | Authors: *Nasrin Mostafazadeh, Nathanael Chambers, Xiaodong He, Devi Parikh, Dhruv Batra, Lucy Vanderwende, Pushmeet Kohli, James Allen* 65 | 66 | Links: [paper](https://www.aclweb.org/anthology/N16-1098/) / [data](https://www.cs.rochester.edu/nlp/rocstories/) 67 | 68 | Type: Cloze; 69 | 70 | 1. [**NarrativeQA**] The NarrativeQA Reading Comprehension Challenge. TACL,2018. 71 | 72 | Authors: *Tomáš Kočiský, Jonathan Schwarz, Phil Blunsom, Chris Dyer, Karl Moritz Hermann, Gábor Melis, Edward Grefenstette* 73 | 74 | Links: [paper](https://arxiv.org/abs/1712.07040) / [data](https://github.com/deepmind/narrativeqa) 75 | 76 | Type: Generation; 77 | 78 | 1. [**SemEval-2018 Task 11**] MCScript: A Novel Dataset for Assessing Machine Comprehension Using Script Knowledge. LERC,2018. 79 | 80 | Authors: *Simon Ostermann, Ashutosh Modi, Michael Roth, Stefan Thater, Manfred Pinkal* 81 | 82 | Links: [paper](https://arxiv.org/abs/1803.05223) / [data](http://www.sfb1102.uni-saarland.de/?page_id=2582) 83 | 84 | Type: Multiple-Choice; 85 | 86 | 1. [**story-commonsense**] Modeling Naive Psychology of Characters in Simple Commonsense Stories. ACL,2018. 87 | 88 | Authors: *Hannah Rashkin, Antoine Bosselut, Maarten Sap, Kevin Knight, Yejin Choi* 89 | 90 | Links: [paper](https://www.aclweb.org/anthology/P18-1213/) / [data](http://uwnlp.github.io/storycommonsense) 91 | 92 | Type: Multiple-Choice; 93 | 94 | 1. **Event2Mind**: Commonsense Inference on Events, Intents, and Reactions. ACL,2018. 95 | 96 | Authors: *Hannah Rashkin, Maarten Sap, Emily Allaway, Noah A. Smith, Yejin Choi* 97 | 98 | Links: [paper](https://www.aclweb.org/anthology/P18-1043/) / [data](https://uwnlp.github.io/event2mind/) 99 | 100 | Types: Generation; 101 | 102 | 1. **ATOMIC**: An Atlas of Machine Commonsense for If-Then Reasoning. AAAI,2019. 103 | 104 | Authors: *Maarten Sap, Ronan LeBras, Emily Allaway, Chandra Bhagavatula, Nicholas Lourie, Hannah Rashkin, Brendan Roof, Noah A. Smith, Yejin Choi* 105 | 106 | Links: [paper](https://homes.cs.washington.edu/~msap/atomic/data/sap2019atomic.pdf) / [data](https://homes.cs.washington.edu/~msap/atomic/) 107 | 108 | Types: Generation; 109 | 110 | 1. [**ARC**] Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge. 2018. 111 | 112 | Authors: *Peter Clark, Isaac Cowhey, Oren Etzioni, Tushar Khot, Ashish Sabharwal, Carissa Schoenick, Oyvind Tafjord* 113 | 114 | Links: [paper](http://ai2-website.s3.amazonaws.com/publications/AI2ReasoningChallenge2018.pdf) / [data](http://data.allenai.org/arc/) 115 | 116 | Type: Multiple-Choice; 117 | 118 | 1. [**OpenBookQA**] Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answering. EMNLP,2018. 119 | 120 | Authors: *Todor Mihaylov, Peter Clark, Tushar Khot, Ashish Sabharwal* 121 | 122 | Links: [paper](https://www.aclweb.org/anthology/D18-1260/) / [data](https://leaderboard.allenai.org/open_book_qa) 123 | 124 | Type: Multiple-Choice; 125 | 126 | 1. **ReCoRD**: Bridging the Gap between Human and Machine Commonsense Reading Comprehension. 2018. 127 | 128 | Authors: *Sheng Zhang, Xiaodong Liu, Jingjing Liu, Jianfeng Gao, Kevin Duh, Benjamin Van Durme* 129 | 130 | Links: [paper](https://arxiv.org/abs/1810.12885) / [data](https://sheng-z.github.io/ReCoRD-explorer/) 131 | 132 | Type: Cloze; 133 | 134 | 1. **CommonsenseQA**: A Question Answering Challenge Targeting Commonsense Knowledge. NAACL,2019. 135 | 136 | Authors: *Alon Talmor, Jonathan Herzig, Nicholas Lourie, Jonathan Berant* 137 | 138 | Links: [paper](https://www.aclweb.org/anthology/N19-1421/) / [data](https://www.tau-nlp.org/commonsenseqa) 139 | 140 | Type: Multiple-Choice; 141 | 142 | 1. **ChID**: A Large-scale Chinese IDiom Dataset for Cloze Test. ACL,2019. 143 | 144 | Authors: *Chujie Zheng, Minlie Huang, Aixin Sun* 145 | 146 | Links: [paper](https://www.aclweb.org/anthology/P19-1075) / [data](https://github.com/chujiezheng/ChID-Dataset) 147 | 148 | Type: Cloze; 149 | 150 | 1. [**sense-making**] Does it Make Sense? And Why? A Pilot Study for Sense Making and Explanation. ACL,2019. 151 | 152 | Authors: *Cunxiang Wang, Shuailong Liang, Yue Zhang, Xiaonan Li, Tian Gao* 153 | 154 | Links: [paper](https://www.aclweb.org/anthology/P19-1393/) / [data]( https://github.com/wangcunxiang/SenMaking-and-Explanation) 155 | 156 | Type: Multiple-Choice; 157 | 158 | 1. **HellaSwag**: Can a Machine Really Finish Your Sentence? ACL,2019. 159 | 160 | Authors: *Rowan Zellers, Ari Holtzman, Yonatan Bisk, Ali Farhadi, Yejin Choi* 161 | 162 | Links: [paper](https://arxiv.org/abs/1905.07830) / [data](https://rowanzellers.com/hellaswag/) 163 | 164 | Type: Multiple-Choice; 165 | 166 | 1. **SocialIQA**: Commonsense Reasoning about Social Interactions. EMNLP,2019. 167 | 168 | Authors: *Maarten Sap, Hannah Rashkin, Derek Chen, Ronan LeBras, Yejin Choi* 169 | 170 | Links: [paper](https://arxiv.org/abs/1904.09728) / [data](https://maartensap.github.io/social-iqa/) 171 | 172 | Type: Multiple-Choice; 173 | 174 | 1. [**ANLI**] Abductive Commonsense Reasoning. 2019. 175 | 176 | Authors: *Chandra Bhagavatula, Ronan Le Bras, Chaitanya Malaviya, Keisuke Sakaguchi, Ari Holtzman, Hannah Rashkin, Doug Downey, Scott Wen-tau Yih, Yejin Choi* 177 | 178 | Links: [paper](https://arxiv.org/abs/1908.05739) / [data](https://leaderboard.allenai.org/anli/submissions/get-started) 179 | 180 | Type: Multiple-Choice; 181 | 182 | 1. **Cosmos QA**: Machine Reading Comprehension with Contextual Commonsense Reasoning. EMNLP,2019. 183 | 184 | Authors: *Lifu Huang, Ronan Le Bras, Chandra Bhagavatula, Yejin Choi* 185 | 186 | Links: [paper](https://arxiv.org/abs/1909.00277) / [data](https://wilburone.github.io/cosmos/) 187 | 188 | Type: Multiple-Choice; 189 | 190 | Note: *Only consider the benchmark datasets/tasks which require knowledge to complete.* 191 | 192 | ## [Cloze Style Tasks](#content) 193 | 194 | 1. **World Knowledge for Reading Comprehension: Rare Entity Prediction with Hierarchical LSTMs Using External Descriptions**. EMNLP,2017. 195 | 196 | Authors: *Teng Long, Emmanuel Bengio, Ryan Lowe Jackie Chi Kit Cheung, Doina Precup* 197 | 198 | Links: [paper](https://www.aclweb.org/anthology/D17-1086) 199 | 200 | Tasks: Rare Entity Prediction; 201 | 202 | 1. **Reasoning with Heterogeneous Knowledge for Commonsense Machine Comprehension**. ACL,2017. 203 | 204 | Authors: *Hongyu Lin, Le Sun, Xianpei Han* 205 | 206 | Links: [paper](http://aclweb.org/anthology/D17-1216) 207 | 208 | Tasks: SCT; 209 | 210 | 1. **Knowledgeable Reader: Enhancing Cloze-Style Reading Comprehension with External Commonsense Knowledge**. ACL,2018. 211 | 212 | Authors: *Todor Mihaylov, Anette Frank* 213 | 214 | Links: [paper](http://aclweb.org/anthology/P18-1076) / code / [note](http://xingluxi.github.io/2019/01/09/paper-knreader/) 215 | 216 | Tasks: Common Nouns; 217 | 218 | 1. **A Multi-Attention based Neural Network with External Knowledge for Story Ending Predicting Task**. COLING,2018. 219 | 220 | Authors: *Qian Li, Ziwei Li, Jin-Mao Wei, Yanhui Gu, Adam Jatowt, Zhenglu Yang* 221 | 222 | Links: [paper](http://www.aclweb.org/anthology/C18-1149) 223 | 224 | Tasks: SCT; 225 | 226 | 1. **Incorporating Structured Commonsense Knowledge in Story Completion**. AAAI,2018. 227 | 228 | Authors: *Jiaao Chen, Jianshu Chen, Zhou Yu* 229 | 230 | Links: [paper](https://arxiv.org/abs/1811.00625) 231 | 232 | Tasks: SCT; 233 | 234 | 2. **Story Ending Prediction by Transferable BERT**. IJCAI,2019. 235 | 236 | Authors: *Zhongyang Li, Xiao Ding, Ting Liu* 237 | 238 | Links: [paper](https://arxiv.org/abs/1905.07504) 239 | 240 | Tasks: SCT; 241 | 242 | 243 | ## [Span Extraction Tasks](#content) 244 | 245 | 1. **Dynamic Integration of Background Knowledge in Neural NLU Systems**. arxiv,2018. 246 | 247 | Authors: *Dirk Weissenborn, Tomas Kocisky, Chris Dyer* 248 | 249 | Links: [paper](https://arxiv.org/pdf/1706.02596.pdf) 250 | 251 | Tasks: SQuAD; TriviaQA; 252 | 253 | 1. **Explicit Utilization of General Knowledge in Machine Reading Comprehension**. ACL,2019. 254 | 255 | Authors: *Chao Wang, Hui Jiang* 256 | 257 | Links: [paper](https://www.aclweb.org/anthology/P19-1219) / [note](http://xingluxi.github.io/2019/07/17/paper-acl2019-kar/) 258 | 259 | Tasks: SQuAD; 260 | 261 | 1. **Enhancing Pre-Trained Language Representations with Rich Knowledge for Machine Reading Comprehension**. ACL,2019. 262 | 263 | Authors: *An Yang, Quan Wang, Jing Liu, Kai Liu, Yajuan Lyu, Hua Wu, Qiaoqiao She, Sujian Li* 264 | 265 | Links: [paper](https://www.aclweb.org/anthology/P19-1226/) / [note](http://xingluxi.github.io/2019/07/29/paper-acl2019-kt-net/) 266 | 267 | Tasks: SQuAD; ReCoRD; 268 | 269 | ## [Multiple Choice Tasks](#content) 270 | 271 | 1. **Yuanfudao at SemEval-2018 Task 11: Three-way Attention and Relational Knowledge for Commonsense Machine Comprehension**. SemEval,2018. 272 | 273 | Author: *Liang Wang, Meng Sun, Wei Zhao, Kewei Shen, Jingming Liu* 274 | 275 | Links: [paper](https://www.aclweb.org/anthology/S18-1120/) / [code](https://github.com/intfloat/commonsense-rc) 276 | 277 | Tasks: SemEval-2018 Task 11; 278 | 279 | 2. **Improving Question Answering by Commonsense-Based Pre-Training**. AAAI,2019. 280 | 281 | Authors: *Wanjun Zhong, Duyu Tang, Nan Duan, Ming Zhou, Jiahai Wang, Jian Yin* 282 | 283 | Links: [paper](https://arxiv.org/abs/1809.03568) 284 | 285 | Tasks: ARC; OpenBookQA; SemEval-2018 Task 11; 286 | 287 | 1. **Improving Machine Reading Comprehension with General Reading Strategies**. NAACL,2019. 288 | 289 | Authors: *Kai Sun, Dian Yu, Dong Yu, Claire Cardie* 290 | 291 | Links: [paper](https://www.aclweb.org/anthology/N19-1270/) / [code](https://github.com/nlpdata/strategy/) 292 | 293 | Tasks: ARC; OpenBookQA; MCTest; SemEval-2018 Task 11; SCT; MultiRC; 294 | 295 | 1. **Ranking and Selecting Multi-Hop Knowledge Paths to Better Predict Human Needs**. NAACL,2019. 296 | 297 | Authors: *Debjit Paul, Anette Frank* 298 | 299 | Links: [paper](https://www.aclweb.org/anthology/papers/N/N19/N19-1368/) / [code](https://github.com/debjitpaul/Multi-Hop-Knowledge-Paths-Human-Needs) 300 | 301 | Tasks: story commonsense; 302 | 303 | 1. **Incorporating Relation Knowledge into Commonsense Reading Comprehension with Multi-task Learning**. CIKM,2019. 304 | 305 | Author: *Jiangnan Xia, Chen Wu, Ming Yan* 306 | 307 | Links: [paper](https://arxiv.org/abs/1908.04530) 308 | 309 | Tasks: SemEval-2018 Task 11 / SCT; 310 | 311 | 1. **Explain Yourself! Leveraging Language Models for Commonsense Reasoning**. ACL,2019. 312 | 313 | Authors: *Nazneen Fatema Rajani, Bryan McCann, Caiming Xiong, Richard Socher* 314 | 315 | Links: [paper](https://www.aclweb.org/anthology/P19-1487.pdf) / [code](https://github.com/salesforce/cos-e) / [note](http://xingluxi.github.io/2019/07/10/paper-acl2019-cos-e/) 316 | 317 | Tasks: CommonsenseQA; 318 | 319 | 1. **Careful Selection of Knowledge to solve Open Book Question Answering**. ACL,2019. 320 | 321 | Authors: *Pratyay Banerjee, Kuntal Kumar Pal, Arindam Mitra, Chitta Baral* 322 | 323 | Links: [paper](https://arxiv.org/abs/1907.10738) 324 | 325 | Tasks: OpenBookQA; 326 | 327 | 1. **Improving Question Answering with External Knowledge**. EMNLP-MRQA,2019. 328 | 329 | Authors: *Xiaoman Pan, Kai Sun, Dian Yu, Jianshu Chen, Heng Ji, Claire Cardie, Dong Yu* 330 | 331 | Links: [paper](https://arxiv.org/pdf/1902.00993.pdf) / [note](http://xingluxi.github.io/2019/08/26/paper-2019-edl-md/) 332 | 333 | Tasks: ARC; OpenBookQA; 334 | 335 | 1. **KagNet: Knowledge-Aware Graph Networks for Commonsense Reasoning**. EMNLP,2019. 336 | 337 | Authors: *Bill Yuchen Lin, Xinyue Chen, Jamin Chen, Xiang Ren* 338 | 339 | Links: [paper](https://arxiv.org/abs/1909.02151) / [code](https://github.com/INK-USC/KagNet) / [note](https://zhuanlan.zhihu.com/p/81917730) 340 | 341 | Tasks: CommonsenseQA; 342 | 343 | 1. **What’s Missing: A Knowledge Gap Guided Approach for Multi-hop Question Answering**. EMNLP,2019. 344 | 345 | Authors: *Tushar Khot, Ashish Sabharwal, Peter Clark* 346 | 347 | Links: [paper](https://arxiv.org/abs/1909.09253)/ [note](https://zhuanlan.zhihu.com/p/85852386) 348 | 349 | Tasks: OpenBookQA; 350 | 351 | 1. **BIG MOOD: Relating Transformers to Explicit Commonsense Knowledge**. EMNLP-COIN,2019. 352 | 353 | Authors: *Jeff Da* 354 | 355 | Links: [paper](http://arxiv.org/abs/1910.07713) 356 | 357 | Tasks: MCScripts-v2; 358 | 359 | ## [Generation Tasks](#content) 360 | Also known as **Free-form Answer Tasks** 361 | 362 | 1. **Commonsense for Generative Multi-Hop Question Answering Tasks**. EMNLP,2018. 363 | 364 | Authors: *Lisa Bauer, Yicheng Wang, Mohit Bansal* 365 | 366 | Links: [paper](https://www.aclweb.org/anthology/D18-1454/) / [code](https://github.com/yicheng-w/CommonSenseMultiHopQA) / [note](http://xingluxi.github.io/2019/02/21/paper-emnlp2018-mhpgm/) 367 | 368 | Tasks: NarrativeQA; QAngaroo-WikiHop; 369 | 370 | 1. **COMET: Commonsense Transformers for Automatic Knowledge Graph Construction**. ACL,2019. 371 | 372 | Authors: *Antoine Bosselut, Hannah Rashkin, Maarten Sap, Chaitanya Malaviya, Asli Celikyilmaz, Yejin Choi* 373 | 374 | Links: [paper](https://arxiv.org/abs/1906.05317) / [code](https://github.com/atcbosselut/comet-commonsense) / [note](https://indexfziq.github.io/2019/07/03/COMET/) 375 | 376 | Tasks: Atomic; 377 | 378 | 1. **Incorporating External Knowledge into Machine Reading for Generative Question Answering**. EMNLP,2019. 379 | 380 | Authors: *Bin Bi, Chen Wu, Ming Yan, Wei Wang, Jiangnan Xia, Chenliang Li* 381 | 382 | Links: [paper](http://arxiv.org/abs/1909.02745) / [note](http://xingluxi.github.io/2019/10/13/paper-emnlp2019-keag/) 383 | 384 | Tasks: MS MARCO; 385 | 386 | 387 | ## Other Paper List About MRC 388 | 389 | [thunlp/RCPapers](https://github.com/thunlp/RCPapers) -------------------------------------------------------------------------------- /v1.md: -------------------------------------------------------------------------------- 1 | # Papers on Knowledge-based Machine Reading Comprehension. 2 | 3 | A list of recent papers about **Knowledge-based Machine Reading Comprehension** (**KMRC**). 4 | 5 | Contributed by [Luxi Xing](https://github.com/XingLuxi) and [Yuqiang Xie](https://github.com/IndexFziQ). 6 | 7 | Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China. 8 | 9 | Update on **Oct. 23, 2019**. 10 | 11 | (the current version only contains the works published on the conferences or journals, we will continuously update this list.) 12 | 13 | ------- 14 | 15 | ## [Content](#content) 16 | 17 | 1. [Survey](#survey-papers) 18 | 2. [Cloze Style Tasks](#cloze-style-tasks) 19 | 3. [Span Extraction Tasks](#span-extraction-tasks) 20 | 4. [Multiple Choice Tasks](#multiple-choice-tasks) 21 | 5. [Generation Tasks](#generation-tasks) 22 | 6. [Datasets](#benchmark-datasets) 23 | 24 | Note: papers about KBQA will be not included in this list. 25 | 26 | ## [Survey Papers](#content) 27 | 28 | 1. **Commonsense Reasoning for Natural Language Understanding: A Survey of Benchmarks, Resources, and Approaches**. 2019. [[paper](https://arxiv.org/abs/1904.01172) / [note](https://zhuanlan.zhihu.com/p/80883568)] 29 | 30 | Authors: *Shane Storks, Qianzi Gao, Joyce Y. Chai* 31 | 32 | 1. **Neural Machine Reading Comprehension: Methods and Trends**. 2019. [[paper](https://arxiv.org/abs/1907.01118)] 33 | 34 | Authors: *Shanshan Liu, Xin Zhang, Sheng Zhang, Hui Wang, Weiming Zhang* 35 | 36 | 1. **Machine Reading Comprehension: a Literature Review**. 2019. [[paper](https://arxiv.org/abs/1907.01686)] 37 | 38 | Authors: *Xin Zhang, An Yang, Sujian Li, Yizhong Wang* 39 | 40 | 41 | ## [Cloze Style Tasks](#content) 42 | 43 | 1. **World Knowledge for Reading Comprehension: Rare Entity Prediction with Hierarchical LSTMs Using External Descriptions**. EMNLP,2017. 44 | 45 | Authors: *Teng Long, Emmanuel Bengio, Ryan Lowe Jackie Chi Kit Cheung, Doina Precup* 46 | 47 | Links: [paper](https://www.aclweb.org/anthology/D17-1086) 48 | 49 | Tasks: Rare Entity Prediction; 50 | 51 | 1. **Reasoning with Heterogeneous Knowledge for Commonsense Machine Comprehension**. ACL,2017. 52 | 53 | Authors: *Hongyu Lin, Le Sun, Xianpei Han* 54 | 55 | Links: [paper](http://aclweb.org/anthology/D17-1216) 56 | 57 | Tasks: SCT; 58 | 59 | 1. **Knowledgeable Reader: Enhancing Cloze-Style Reading Comprehension with External Commonsense Knowledge**. ACL,2018. 60 | 61 | Authors: *Todor Mihaylov, Anette Frank* 62 | 63 | Links: [paper](http://aclweb.org/anthology/P18-1076) / code / [note](http://xingluxi.github.io/2019/01/09/paper-knreader/) 64 | 65 | Tasks: Common Nouns; 66 | 67 | 1. **A Multi-Attention based Neural Network with External Knowledge for Story Ending Predicting Task**. COLING,2018. 68 | 69 | Authors: *Qian Li, Ziwei Li, Jin-Mao Wei, Yanhui Gu, Adam Jatowt, Zhenglu Yang* 70 | 71 | Links: [paper](http://www.aclweb.org/anthology/C18-1149) 72 | 73 | Tasks: SCT; 74 | 75 | 1. **Incorporating Structured Commonsense Knowledge in Story Completion**. AAAI,2018. 76 | 77 | Authors: *Jiaao Chen, Jianshu Chen, Zhou Yu* 78 | 79 | Links: [paper](https://arxiv.org/abs/1811.00625) 80 | 81 | Tasks: SCT; 82 | 83 | 2. **Story Ending Prediction by Transferable BERT**. IJCAI,2019. 84 | 85 | Authors: *Zhongyang Li, Xiao Ding, Ting Liu* 86 | 87 | Links: [paper](https://arxiv.org/abs/1905.07504) 88 | 89 | Tasks: SCT; 90 | 91 | 92 | ## [Span Extraction Tasks](#content) 93 | 94 | 1. **Dynamic Integration of Background Knowledge in Neural NLU Systems**. arxiv,2018. 95 | 96 | Authors: *Dirk Weissenborn, Tomas Kocisky, Chris Dyer* 97 | 98 | Links: [paper](https://arxiv.org/pdf/1706.02596.pdf) 99 | 100 | Tasks: SQuAD; TriviaQA; 101 | 102 | 1. **Explicit Utilization of General Knowledge in Machine Reading Comprehension**. ACL,2019. 103 | 104 | Authors: *Chao Wang, Hui Jiang* 105 | 106 | Links: [paper](https://www.aclweb.org/anthology/P19-1219) / [note](http://xingluxi.github.io/2019/07/17/paper-acl2019-kar/) 107 | 108 | Tasks: SQuAD; 109 | 110 | 1. **Enhancing Pre-Trained Language Representations with Rich Knowledge for Machine Reading Comprehension**. ACL,2019. 111 | 112 | Authors: *An Yang, Quan Wang, Jing Liu, Kai Liu, Yajuan Lyu, Hua Wu, Qiaoqiao She, Sujian Li* 113 | 114 | Links: [paper](https://www.aclweb.org/anthology/P19-1226/) / [note](http://xingluxi.github.io/2019/07/29/paper-acl2019-kt-net/) 115 | 116 | Tasks: SQuAD; ReCoRD; 117 | 118 | ## [Multiple Choice Tasks](#content) 119 | 120 | 1. **Yuanfudao at SemEval-2018 Task 11: Three-way Attention and Relational Knowledge for Commonsense Machine Comprehension**. SemEval,2018. 121 | 122 | Author: *Liang Wang, Meng Sun, Wei Zhao, Kewei Shen, Jingming Liu* 123 | 124 | Links: [paper](https://www.aclweb.org/anthology/S18-1120/) / [code](https://github.com/intfloat/commonsense-rc) 125 | 126 | Tasks: SemEval-2018 Task 11; 127 | 128 | 2. **Improving Question Answering by Commonsense-Based Pre-Training**. AAAI,2019. 129 | 130 | Authors: *Wanjun Zhong, Duyu Tang, Nan Duan, Ming Zhou, Jiahai Wang, Jian Yin* 131 | 132 | Links: [paper](https://arxiv.org/abs/1809.03568) 133 | 134 | Tasks: ARC; OpenBookQA; SemEval-2018 Task 11; 135 | 136 | 1. **Improving Machine Reading Comprehension with General Reading Strategies**. NAACL,2019. 137 | 138 | Authors: *Kai Sun, Dian Yu, Dong Yu, Claire Cardie* 139 | 140 | Links: [paper](https://www.aclweb.org/anthology/N19-1270/) / [code](https://github.com/nlpdata/strategy/) 141 | 142 | Tasks: ARC; OpenBookQA; MCTest; SemEval-2018 Task 11; SCT; MultiRC; 143 | 144 | 1. **Ranking and Selecting Multi-Hop Knowledge Paths to Better Predict Human Needs**. NAACL,2019. 145 | 146 | Authors: *Debjit Paul, Anette Frank* 147 | 148 | Links: [paper](https://www.aclweb.org/anthology/papers/N/N19/N19-1368/) / [code](https://github.com/debjitpaul/Multi-Hop-Knowledge-Paths-Human-Needs) 149 | 150 | Tasks: story commonsense; 151 | 152 | 1. **Incorporating Relation Knowledge into Commonsense Reading Comprehension with Multi-task Learning**. CIKM,2019. 153 | 154 | Author: *Jiangnan Xia, Chen Wu, Ming Yan* 155 | 156 | Links: [paper](https://arxiv.org/abs/1908.04530) 157 | 158 | Tasks: SemEval-2018 Task 11 / SCT; 159 | 160 | 1. **Explain Yourself! Leveraging Language Models for Commonsense Reasoning**. ACL,2019. 161 | 162 | Authors: *Nazneen Fatema Rajani, Bryan McCann, Caiming Xiong, Richard Socher* 163 | 164 | Links: [paper](https://www.aclweb.org/anthology/P19-1487.pdf) / [code](https://github.com/salesforce/cos-e) / [note](http://xingluxi.github.io/2019/07/10/paper-acl2019-cos-e/) 165 | 166 | Tasks: CommonsenseQA; 167 | 168 | 1. **Careful Selection of Knowledge to solve Open Book Question Answering**. ACL,2019. 169 | 170 | Authors: *Pratyay Banerjee, Kuntal Kumar Pal, Arindam Mitra, Chitta Baral* 171 | 172 | Links: [paper](https://arxiv.org/abs/1907.10738) 173 | 174 | Tasks: OpenBookQA; 175 | 176 | 1. **Improving Question Answering with External Knowledge**. EMNLP-MRQA,2019. 177 | 178 | Authors: *Xiaoman Pan, Kai Sun, Dian Yu, Jianshu Chen, Heng Ji, Claire Cardie, Dong Yu* 179 | 180 | Links: [paper](https://arxiv.org/pdf/1902.00993.pdf) / [note](http://xingluxi.github.io/2019/08/26/paper-2019-edl-md/) 181 | 182 | Tasks: ARC; OpenBookQA; 183 | 184 | 1. **KagNet: Knowledge-Aware Graph Networks for Commonsense Reasoning**. EMNLP,2019. 185 | 186 | Authors: *Bill Yuchen Lin, Xinyue Chen, Jamin Chen, Xiang Ren* 187 | 188 | Links: [paper](https://arxiv.org/abs/1909.02151) / [code](https://github.com/INK-USC/KagNet) / [note](https://zhuanlan.zhihu.com/p/81917730) 189 | 190 | Tasks: CommonsenseQA; 191 | 192 | 1. **What’s Missing: A Knowledge Gap Guided Approach for Multi-hop Question Answering**. EMNLP,2019. 193 | 194 | Authors: *Tushar Khot, Ashish Sabharwal, Peter Clark* 195 | 196 | Links: [paper](https://arxiv.org/abs/1909.09253)/ [note](https://zhuanlan.zhihu.com/p/85852386) 197 | 198 | Tasks: OpenBookQA; 199 | 200 | 1. **BIG MOOD: Relating Transformers to Explicit Commonsense Knowledge**. EMNLP-COIN,2019. 201 | 202 | Authors: *Jeff Da* 203 | 204 | Links: [paper](http://arxiv.org/abs/1910.07713) 205 | 206 | Tasks: MCScripts-v2; 207 | 208 | ## [Generation Tasks](#content) 209 | Also known as **Free-form Answer Tasks** 210 | 211 | 1. **Commonsense for Generative Multi-Hop Question Answering Tasks**. EMNLP,2018. 212 | 213 | Authors: *Lisa Bauer, Yicheng Wang, Mohit Bansal* 214 | 215 | Links: [paper](https://www.aclweb.org/anthology/D18-1454/) / [code](https://github.com/yicheng-w/CommonSenseMultiHopQA) / [note](http://xingluxi.github.io/2019/02/21/paper-emnlp2018-mhpgm/) 216 | 217 | Tasks: NarrativeQA; QAngaroo-WikiHop; 218 | 219 | 1. **COMET: Commonsense Transformers for Automatic Knowledge Graph Construction**. ACL,2019. 220 | 221 | Authors: *Antoine Bosselut, Hannah Rashkin, Maarten Sap, Chaitanya Malaviya, Asli Celikyilmaz, Yejin Choi* 222 | 223 | Links: [paper](https://arxiv.org/abs/1906.05317) / [code](https://github.com/atcbosselut/comet-commonsense) / [note](https://indexfziq.github.io/2019/07/03/COMET/) 224 | 225 | Tasks: Atomic; 226 | 227 | 1. **Incorporating External Knowledge into Machine Reading for Generative Question Answering**. EMNLP,2019. 228 | 229 | Authors: *Bin Bi, Chen Wu, Ming Yan, Wei Wang, Jiangnan Xia, Chenliang Li* 230 | 231 | Links: [paper](http://arxiv.org/abs/1909.02745) / [note](http://xingluxi.github.io/2019/10/13/paper-emnlp2019-keag/) 232 | 233 | Tasks: MS MARCO; 234 | 235 | 236 | ## [Benchmark Datasets](#content) 237 | 238 | 1. [**COPA**] Choice of Plausible Alternatives: An Evaluation of Commonsense Causal Reasoning. AAAI,2011. [[paper](http://ict.usc.edu/~gordon/publications/AAAI-SPRING11A.PDF) / [data](http://people.ict.usc.edu/~gordon/copa.html)] 239 | 240 | Authors: *Melissa Roemmele, Cosmin Adrian Bejan, Andrew S. Gordon* 241 | * Type: Multiple-Choice; 242 | 243 | 1. [**WSC**] The Winograd Schema Challenge. AAAI,2011. [[paper](https://www.aaai.org/ocs/index.php/KR/KR12/paper/download/4492/4924) /[data](https://cs.nyu.edu/faculty/davise/papers/WinogradSchemas/WS.html)] 244 | 245 | Authors: *Hector J. Levesque, Ernest Davis, Leora Morgenstern* 246 | * Type: Multiple-Choice; 247 | 248 | 2. [**ROCStories**; **SCT**] A Corpus and Cloze Evaluation for Deeper Understanding of Commonsense Stories. NAACL,2016. [[paper](https://www.aclweb.org/anthology/N16-1098/) / [data](https://www.cs.rochester.edu/nlp/rocstories/)] 249 | 250 | Authors: *Nasrin Mostafazadeh, Nathanael Chambers, Xiaodong He, Devi Parikh, Dhruv Batra, Lucy Vanderwende, Pushmeet Kohli, James Allen* 251 | * Type: Cloze; 252 | 253 | 1. [**NarrativeQA**] The NarrativeQA Reading Comprehension Challenge. TACL,2018. [[paper](https://arxiv.org/abs/1712.07040) / [data](https://github.com/deepmind/narrativeqa)] 254 | 255 | Authors: *Tomáš Kočiský, Jonathan Schwarz, Phil Blunsom, Chris Dyer, Karl Moritz Hermann, Gábor Melis, Edward Grefenstette* 256 | * Type: Generation; 257 | 258 | 1. [**SemEval-2018 Task 11**] MCScript: A Novel Dataset for Assessing Machine Comprehension Using Script Knowledge. LERC,2018. [[paper](https://arxiv.org/abs/1803.05223) / [data](http://www.sfb1102.uni-saarland.de/?page_id=2582)] 259 | 260 | Authors: *Simon Ostermann, Ashutosh Modi, Michael Roth, Stefan Thater, Manfred Pinkal* 261 | 262 | * Type: Multiple-Choice; 263 | 264 | 1. [**story-commonsense**] Modeling Naive Psychology of Characters in Simple Commonsense Stories. ACL,2018. [[paper](https://www.aclweb.org/anthology/P18-1213/) / [data](http://uwnlp.github.io/storycommonsense)] 265 | 266 | Authors: *Hannah Rashkin, Antoine Bosselut, Maarten Sap, Kevin Knight, Yejin Choi* 267 | 268 | * Type: Multiple-Choice; 269 | 270 | 1. **Event2Mind**: Commonsense Inference on Events, Intents, and Reactions. ACL,2018. [[paper](https://www.aclweb.org/anthology/P18-1043/) / [data](https://uwnlp.github.io/event2mind/)] 271 | 272 | Authors: *Hannah Rashkin, Maarten Sap, Emily Allaway, Noah A. Smith, Yejin Choi* 273 | 274 | * Types: Generation; 275 | 276 | 1. **ATOMIC**: An Atlas of Machine Commonsense for If-Then Reasoning. AAAI,2019. [[paper](https://homes.cs.washington.edu/~msap/atomic/data/sap2019atomic.pdf) / [data](https://homes.cs.washington.edu/~msap/atomic/)] 277 | 278 | Authors: *Maarten Sap, Ronan LeBras, Emily Allaway, Chandra Bhagavatula, Nicholas Lourie, Hannah Rashkin, Brendan Roof, Noah A. Smith, Yejin Choi* 279 | 280 | * Types: Generation; 281 | 282 | 1. [**ARC**] Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge. 2018. [[paper](http://ai2-website.s3.amazonaws.com/publications/AI2ReasoningChallenge2018.pdf) / [data](http://data.allenai.org/arc/)] 283 | 284 | Authors: *Peter Clark, Isaac Cowhey, Oren Etzioni, Tushar Khot, Ashish Sabharwal, Carissa Schoenick, Oyvind Tafjord* 285 | 286 | * Type: Multiple-Choice; 287 | 288 | 1. [**OpenBookQA**] Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answering. EMNLP,2018. [[paper](https://www.aclweb.org/anthology/D18-1260/) / [data](https://leaderboard.allenai.org/open_book_qa)] 289 | 290 | Authors: *Todor Mihaylov, Peter Clark, Tushar Khot, Ashish Sabharwal* 291 | 292 | * Type: Multiple-Choice; 293 | 294 | 1. **ReCoRD**: Bridging the Gap between Human and Machine Commonsense Reading Comprehension. 2018. [[paper](https://arxiv.org/abs/1810.12885) / [data](https://sheng-z.github.io/ReCoRD-explorer/)] 295 | 296 | Authors: *Sheng Zhang, Xiaodong Liu, Jingjing Liu, Jianfeng Gao, Kevin Duh, Benjamin Van Durme* 297 | 298 | * Type: Cloze; 299 | 300 | 1. **CommonsenseQA**: A Question Answering Challenge Targeting Commonsense Knowledge. NAACL,2019. [[paper](https://www.aclweb.org/anthology/N19-1421/) / [data](https://www.tau-nlp.org/commonsenseqa)] 301 | 302 | Authors: *Alon Talmor, Jonathan Herzig, Nicholas Lourie, Jonathan Berant* 303 | 304 | * Type: Multiple-Choice; 305 | 306 | 1. **ChID**: A Large-scale Chinese IDiom Dataset for Cloze Test. ACL,2019. [[paper](https://www.aclweb.org/anthology/P19-1075) / [data](https://github.com/chujiezheng/ChID-Dataset)] 307 | 308 | Authors: *Chujie Zheng, Minlie Huang, Aixin Sun* 309 | 310 | * Type: Cloze; 311 | 312 | 1. [**sense-making**] Does it Make Sense? And Why? A Pilot Study for Sense Making and Explanation. ACL,2019. [[paper](https://www.aclweb.org/anthology/P19-1393/) / [data]( https://github.com/wangcunxiang/SenMaking-and-Explanation)] 313 | 314 | Authors: *Cunxiang Wang, Shuailong Liang, Yue Zhang, Xiaonan Li, Tian Gao* 315 | 316 | * Type: Multiple-Choice; 317 | 318 | 1. **HellaSwag**: Can a Machine Really Finish Your Sentence? ACL,2019. [[paper](https://arxiv.org/abs/1905.07830) / [data](https://rowanzellers.com/hellaswag/)] 319 | 320 | Authors: *Rowan Zellers, Ari Holtzman, Yonatan Bisk, Ali Farhadi, Yejin Choi* 321 | 322 | * Type: Multiple-Choice; 323 | 324 | 1. **SocialIQA**: Commonsense Reasoning about Social Interactions. EMNLP,2019. [[paper](https://arxiv.org/abs/1904.09728) / [data](https://maartensap.github.io/social-iqa/)] 325 | 326 | Authors: *Maarten Sap, Hannah Rashkin, Derek Chen, Ronan LeBras, Yejin Choi* 327 | 328 | * Type: Multiple-Choice; 329 | 330 | 1. [**ANLI**] Abductive Commonsense Reasoning. 2019. [[paper](https://arxiv.org/abs/1908.05739) / [data](https://leaderboard.allenai.org/anli/submissions/get-started)] 331 | 332 | Authors: *Chandra Bhagavatula, Ronan Le Bras, Chaitanya Malaviya, Keisuke Sakaguchi, Ari Holtzman, Hannah Rashkin, Doug Downey, Scott Wen-tau Yih, Yejin Choi* 333 | 334 | * Type: Multiple-Choice; 335 | 336 | 1. **Cosmos QA**: Machine Reading Comprehension with Contextual Commonsense Reasoning. EMNLP,2019. [[paper](https://arxiv.org/abs/1909.00277) / [data](https://wilburone.github.io/cosmos/)] 337 | 338 | Authors: *Lifu Huang, Ronan Le Bras, Chandra Bhagavatula, Yejin Choi* 339 | 340 | * Type: Multiple-Choice; 341 | 342 | Note: *Only consider the benchmark datasets/tasks which require knowledge to complete.* 343 | 344 | 345 | ## Other Paper List About MRC 346 | 347 | [thunlp/RCPapers](https://github.com/thunlp/RCPapers) 348 | --------------------------------------------------------------------------------