Junbei Zhang, Xiaodan Zhu, Qian Chen, Li-Rong Dai, Si Wei, Hui Jiang
The last several years have seen intensive interest in exploring neural-network-based models for machine comprehension (MC) and question answering (QA). In this paper, we approach the problems by closely modelling questions in a neural network framework. We first introduce syntactic information to help encode questions. We then view and model different types of questions and the information shared among them as an adaptation task and proposed adaptation models for them. On the Stanford Question Answering Dataset (SQuAD), we show that these approaches can help attain better results over a competitive baseline.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Question Answering | SQuAD1.1 dev | EM | 69.1 | jNet (TreeLSTM adaptation, QTLa, K=100) |
| Question Answering | SQuAD1.1 dev | F1 | 78.38 | jNet (TreeLSTM adaptation, QTLa, K=100) |
| Question Answering | SQuAD1.1 | EM | 73.01 | jNet (ensemble) |
| Question Answering | SQuAD1.1 | F1 | 81.517 | jNet (ensemble) |
| Question Answering | SQuAD1.1 | EM | 70.607 | jNet (single model) |
| Question Answering | SQuAD1.1 | F1 | 79.821 | jNet (single model) |