近幾年問題生成的研究發展迅速，過去以句子的語法結構定義規則生成問題，隨著深度學習成熟的技術，現今機器能理解語意並自動產生適當的問題。

問題生成的目標是給定一段文字訊息與答案，產生相對應的問題，與機器閱讀理解任務類似，因此閱讀理解的資料集常被用在問題生成任務中。以往問題生成模型的輸入並非整篇文章，而是包含答案的句子，但有些問題的內容和答案不在同一個句子，可能是依據其他句子資訊產生該答案的問題，於是本論文提出一個新架構，由句對模型和問題生成模型所組成，利用句對模型處理文章結構，將每一句資訊與包含答案的句子進行匹配，計算各自的相關程度並且重新賦予句子權重，接著傳送到問題生成模型產生最終的問題。句對模型主要目的是從整篇文章中自動找尋和答案有關的內容進而產生適合的問題。

實驗結果表示，我們的系統能有效處理文章結構，相比只有問題生成模型的系統，在中文和英文的資料集都有更好的表現。

In recent years, question generation (QG) has developed rapidly. In the past, using rules that are based on syntactic structure to generate questions. Nowadays, the machine can understand semantic and automatically generate appropriate questions with a proven technique of deep learning.

Question generation aims to generate corresponding questions from a given passage and answer. It is similar to machine reading comprehension (RC) task. Therefore, reading comprehension dataset is often used to question generation task. The input of the previous question generation model is the sentence containing the answer rather than the whole article. However, the content of some questions and its answers are not in the same sentence. The question may be based on other information in sentences. Then, our paper proposed a new framework which consists of sentence pair model and question generation model. Using the sentence pair model to process article structure. Its method is matching each sentence and the sentence containing the answer to compute the respective degree of correlation to reweight sentences and then produce questions by question generation model. The main purpose of sentence pair model is to automatically find the content related to the answer from the article.

Experiment results show that our system can handle article structure. In contrast to a system with only question generation model, our system has better performance in Chinese and English dataset.

[1] P. Rajpurkar, J. Zhang, K. Lopyrev, and P. Liang, “Squad: 100,000+ questions for
machine comprehension of text,” arXiv preprint arXiv:1606.05250, 2016.
[2] X. Du, J. Shao, and C. Cardie, “Learning to ask: Neural question generation for
reading comprehension,” arXiv preprint arXiv:1705.00106, 2017.
[3] C. C. Shao, T. Liu, Y. Lai, Y. Tseng, and S. Tsai, “Drcd: a chinese machine reading
comprehension dataset,” arXiv preprint arXiv:1806.00920, 2018.
[4] I. Sutskever, O. Vinyals, and Q. V. Le, “Sequence to sequence learning with neural
networks,” in Advances in neural information processing systems, pp. 3104–3112,
2014.
[5] D. Bahdanau, K. Cho, and Y. Bengio, “Neural machine translation by jointly learning
to align and translate,” arXiv preprint arXiv:1409.0473, 2014.
[6] M. Schuster and K. K. Paliwal, “Bidirectional recurrent neural networks,” IEEE
Transactions on Signal Processing, vol. 45, no. 11, pp. 2673–2681, 1997.
[7] M. Heilman and N. A. Smith, “Good question! statistical ranking for question generation,”
in Human Language Technologies: The 2010 Annual Conference of the
North American Chapter of the Association for Computational Linguistics, pp. 609–
617, Association for Computational Linguistics, 2010.
[8] Q. Zhou, N. Yang, F. Wei, C. Tan, H. Bao, and M. Zhou, “Neural question generation
from text: A preliminary study,” in National CCF Conference on Natural Language
Processing and Chinese Computing, pp. 662–671, Springer, 2017.
[9] C. Gulcehre, S. Ahn, R. Nallapati, B. Zhou, and Y. Bengio, “Pointing the unknown
words,” arXiv preprint arXiv:1603.08148, 2016.
[10] J. Gu, Z. Lu, H. Li, and V. O. Li, “Incorporating copying mechanism in sequenceto-
sequence learning,” arXiv preprint arXiv:1603.06393, 2016.
[11] L. Song, Z. Wang, W. Hamza, Y. Zhang, and D. Gildea, “Leveraging context information
for natural question generation,” in Proceedings of the 2018 Conference
of the North American Chapter of the Association for Computational Linguistics:
Human Language Technologies, Volume 2 (Short Papers), pp. 569–574, 2018.
[12] Z. Wang, H. Mi, W. Hamza, and R. Florian, “Multi-perspective context matching
for machine comprehension,” arXiv preprint arXiv:1612.04211, 2016.
[13] V. Kumar, G. Ramakrishnan, and Y.-F. Li, “A framework for automatic question
generation from text using deep reinforcement learning,” arXiv preprint arXiv:
1808.04961, 2018.
[14] Z. Tu, Z. Lu, Y. Liu, X. Liu, and H. Li, “Modeling coverage for neural machine
translation,” arXiv preprint arXiv:1601.04811, 2016.
[15] K. Papineni, S. Roukos, T. Ward, and W.-J. Zhu, “Bleu: a method for automatic
evaluation of machine translation,” in Proceedings of the 40th annual meeting on
association for computational linguistics, pp. 311–318, Association for Computational
Linguistics, 2002.
[16] C.-Y. Lin, “Rouge: A package for automatic evaluation of summaries,” Text Summarization
Branches Out, 2004.
[17] Y. Kim, H. Lee, J. Shin, and K. Jung, “Improving neural question generation using
answer separation,” arXiv preprint arXiv:1809.02393, 2018.
[18] S. R. Bowman, G. Angeli, C. Potts, and C. D. Manning, “A large annotated corpus
for learning natural language inference,” arXiv preprint arXiv:1508.05326, 2015.
[19] S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural computation,
vol. 9, no. 8, pp. 1735–1780, 1997.
[20] T. Rocktäschel, E. Grefenstette, K. M. Hermann, T. Kočiskỳ, and P. Blunsom, “Reasoning
about entailment with neural attention,” arXiv preprint arXiv:1509.06664,
2015.
[21] W. Yin, H. Schütze, B. Xiang, and B. Zhou, “Abcnn: Attention-based convolutional
neural network for modeling sentence pairs,” Transactions of the Association for
Computational Linguistics, vol. 4, pp. 259–272, 2016.
[22] S. Wang and J. Jiang, “Learning natural language inference with lstm,” arXiv
preprint arXiv:1512.08849, 2015.
[23] T. Mikolov, I. Sutskever, K. Chen, G. S. Corrado, and J. Dean, “Distributed representations
of words and phrases and their compositionality,” in Advances in neural
information processing systems, pp. 3111–3119, 2013.
[24] D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” arXiv
preprint arXiv:1412.6980, 2014.