Yuanfudao at SemEval-2018 Task 11: Three-way Attention and Relational Knowledge for Commonsense Machine Comprehension
arXiv:1803.00191
Abstract
This paper describes our system for SemEval-2018 Task 11: Machine Comprehension using Commonsense Knowledge. We use Three-way Attentive Networks (TriAN) to model interactions between the passage, question and answers. To incorporate commonsense knowledge, we augment the input with relation embedding from the graph of general knowledge ConceptNet (Speer et al., 2017). As a result, our system achieves state-of-the-art performance with 83.95% accuracy on the official test data. Code is publicly available at https://github.com/intfloat/commonsense-rc
arXiv comment5 pages, 1 figure, Accepted to International Workshop on Semantic Evaluation 2018
References in corpus (6)
- Adam: A Method for Stochastic Optimization
- ConceptNet 5.5: An Open Multilingual Graph of General Knowledge
- SQuAD: 100,000+ Questions for Machine Comprehension of Text
- RACE: Large-scale ReAding Comprehension Dataset From Examinations
- Reading Wikipedia to Answer Open-Domain Questions
- MCScript: A Novel Dataset for Assessing Machine Comprehension Using Script Knowledge
Cited by in corpus (9)
- Convolutional Spatial Attention Model for Reading Comprehension with Multiple-Choice Questions
- Improving Question Answering by Commonsense-Based Pre-Training
- Consolidating Commonsense Knowledge
- WinoWhy: A Deep Diagnosis of Essential Commonsense Knowledge for Answering Winograd Schema Challenge
- TransOMCS: From Linguistic Graphs to Commonsense Knowledge
- Inferential Text Generation with Multiple Knowledge Sources and Meta-Learning
- Weighted Global Normalization for Multiple Choice Reading Comprehension over Long Documents
- Understanding in Artificial Intelligence
- BIG MOOD: Relating Transformers to Explicit Commonsense Knowledge