Question-type Driven Question Generation

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 6032–6037, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 6032 Question-type Driven Question Generation Wenjie Zhou, Minghua Zhang, Yunfang Wu∗ Key Laboratory of Computational Linguistics, Ministry of Education School of Electronics Engineering and Computer Science, Peking University, Beijing, China {wjzhou013, zhangmh, wuyf}@pku.edu.cn Abstract Question generation is a challenging task which aims to ask a question based on an an- swer and relevant context. The existing works suffer from the mismatching between question type and answer, i.e. generating a question with type how while the answer is a personal name. We propose to automatically predict the question type based on the input answer and context. Then, the question type is fused into a seq2seq model to guide the question genera- tion, so as to deal with the mismatching prob- lem. We achieve significant improvement on the accuracy of question type prediction and finally obtain state-of-the-art results for ques- tion generation on both SQuAD and MARCO datasets. 1 Introduction Question generation (QG) can be effectively ap- plied to many fields, including question answering (Duan et al., 2017), dialogue system (Shum et al., 2018) and education. In this paper, we focus on the answer-aware QG, which is to generate a question according to the given sentence and the expected answer. Recently, the neural-based approaches on QG have achieved remarkable success, by applying large-scale reading comprehension datasets and employing the encoder-decoder framework. Most of the existing works are based on the seq2seq network incorporating attention mechanism and copy mode, which are first applied in Zhou et al. (2017). Later, Song et al. (2018) lever- age multi-perspective matching methods, and Sun et al. (2018) propose a position-aware model to put more emphasis on answer-surrounded context words. Both works are trying to enhance the rel- evance between the question and answer. Zhao et al. (2018) aggregate paragraph-level context to ∗Corresponding author. provide sufficient information for question gener- ation. Another direction is to integrate question answering and question generation as dual tasks (Tang et al., 2017). Beyond answer-aware QG, some systems try to generate questions from a text without answer as input (Du and Cardie, 2017; Subramanian et al., 2018). Despite the progress achieved by the previous work, we found that the types of generated ques- tions are often incorrect. According to experi- ments on SQuAD, one strong model (Zhou et al., 2017) we replicate only obtains 57.6% accuracy in question type. As we know, question type is vital for question generation, since it determines the question pattern and guides the generating pro- cess. If the question type is incorrect, the remained generated sequence would drift far away. Several works have addressed this issue. Sun et al. (2018) incorporated a question word gener- ation mode to generate question word at each de- coding step, which utilized the answer information by employing the encoder hidden states at the an- swer start position. …

Content truncated to 500 words