Abstract
This study focuses on natural answer generation that generates a complete answer using given paragraphs. Existing methods of only extracting the answer span loses the naturalness while those of generating it word by word increases the learning difficulty. In this paper, we propose to split the answering process into two branches, which share the same encoder and deeply interact with each other. On one branch, we generate the answer template based on the question through text editing method. On the other branch, we extract the answer span based on the documents. Finally, the answer is composed of the generated answer template and the extracted answer span. Besides, we propose to select the span in candidates to better fit the template with more fluency. Experiments show that our method improves the performance on natural answer generation.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
The span is provided in the dataset. If not, we can adopt Rouge-L or Edit Distance to match it in the documents with the natural answer.
- 2.
References
Bajaj, P., et al.: MS MARCO: a human generated machine reading comprehension dataset. arXiv preprint arXiv:1611.09268 (2016)
Beltagy, I., Peters, M.E., Cohan, A.: Longformer: the long-document transformer. arXiv preprint arXiv:2004.05150 (2020)
Cho, K., et al.: Learning phrase representations using RNN encoder-decoder for statistical machine translation. In: EMNLP, pp. 1724–1734 (2014)
Cui, Y., et al.: A span-extraction dataset for Chinese machine reading comprehension. In: EMNLP, pp. 5886–5891 (2019)
Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: Bert: pre-training of deep bidirectional transformers for language understanding. In: NAACL, pp. 4171–4186 (2019)
Dong, Y., Li, Z., Rezagholizadeh, M., Cheung, J.C.K.: EditNTS: an neural programmer-interpreter model for sentence simplification through explicit editing. In: ACL, pp. 3393–3402 (2019)
Filippova, K., Alfonseca, E., Colmenares, C.A., Kaiser, Ł., Vinyals, O.: Sentence compression by deletion with LSTMs. In: EMNLP, pp. 360–368 (2015)
Fu, Y., Feng, Y.: Natural answer generation with heterogeneous memory. In: NAACL, pp. 185–195 (2018)
Ge, T., Wei, F., Zhou, M.: Fluency boost learning and inference for neural grammatical error correction. In: ACL, pp. 1055–1065 (2018)
Gu, J., Lu, Z., Li, H., Li, V.O.: Incorporating copying mechanism in sequence-to-sequence learning. In: ACL, pp. 1631–1640 (2016)
He, S., Liu, C., Liu, K., Zhao, J.: Generating natural answers by incorporating copying and retrieving mechanisms in sequence-to-sequence learning. In: ACL, pp. 199–208 (2017)
He, W., et al.: DuReader: a Chinese machine reading comprehension dataset from real-world applications. In: ACL, pp. 37–46 (2018)
Malmi, E., Krause, S., Rothe, S., Mirylenka, D., Severyn, A.: Encode, tag, realize: High-precision text editing. In: EMNLP, pp. 5054–5065 (2019)
Narayan, S., Cohen, S.B., Lapata, M.: Ranking sentences for extractive summarization with reinforcement learning. In: NAACL, pp. 1747–1759 (2018)
Nishida, K., et al.: Multi-style generative reading comprehension. In: ACL, pp. 2273–2284 (2019)
Rajpurkar, P., Zhang, J., Lopyrev, K., Liang, P.: Squad: 100,000+ questions for machine comprehension of text. In: EMNLP, pp. 2383–2392 (2016)
Seo, M., Kembhavi, A., Farhadi, A., Hajishirzi, H.: Bidirectional attention flow for machine comprehension. arXiv preprint arXiv:1611.01603 (2016)
Tan, C., Wei, F., Yang, N., Du, B., Lv, W., Zhou, M.: S-net: from answer extraction to answer generation for machine reading comprehension. arXiv preprint arXiv:1706.04815 (2017)
Wang, Y., et al.: Multi-passage machine reading comprehension with cross-passage answer verification. In: ACL, pp. 1918–1927 (2018)
Yan, M., et al.: A deep cascade model for multi-document reading comprehension. In: AAAI, pp. 7354–7361 (2019)
Yang, Z., et al.: HotpotQA: a dataset for diverse, explainable multi-hop question answering. In: EMNLP, pp. 2369–2380 (2018)
Yao, Z., Xu, F., Yin, P., Sun, H., Neubig, G.: Learning structural edits via incremental tree transformations. In: ICLR (2021)
Yin, J., Jiang, X., Lu, Z., Shang, L., Li, H., Li, X.: Neural generative question answering. In: IJCAI, pp. 2972–2978 (2016)
Yu, A.W., et al.: QANet: combining local convolution with global self-attention for reading comprehension. In: ICLR (2018)
Zhao, W., Wang, L., Shen, K., Jia, R., Liu, J.: Improving grammatical error correction via pre-training a copy-augmented architecture with unlabeled data. In: NAACL, pp. 156–165 (2019)
Acknowledgement
This work is partially supported by National Natural Science Foundation of China (Grants no. 61772568), Guangdong Basic and Applied Basic Research Foundation (Grant no. 2019A1515012029), and Youth science and technology innovation talent of Guangdong Special Support Program.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Huang, Y., Yang, M., Yang, N. (2021). Generating Relevant, Correct and Fluent Answers in Natural Answer Generation. In: Wang, L., Feng, Y., Hong, Y., He, R. (eds) Natural Language Processing and Chinese Computing. NLPCC 2021. Lecture Notes in Computer Science(), vol 13029. Springer, Cham. https://doi.org/10.1007/978-3-030-88483-3_12
Download citation
DOI: https://doi.org/10.1007/978-3-030-88483-3_12
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-88482-6
Online ISBN: 978-3-030-88483-3
eBook Packages: Computer ScienceComputer Science (R0)
