├── 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 |
--------------------------------------------------------------------------------