Abstract
The retriever-reader framework has been widely used in open-domain question answering with great success. Current studies show that better retrieval can greatly improve the performance of final answer extraction and may replace the reader stage. Considering the limited computing resources and the great progress that has been made in reading comprehension, we continue to use the retriever-reader framework and focus on efficient retrieval. In this paper, we propose a new coarse-to-fine retrieval method to take away the semantic noise left by coarse-grained filtering. In particular, we join a fine-grained retriever after the passages generated by the coarse-grained retriever, making all sentences in the passage match more closely with the question. Meanwhile, we use contrastive learning to construct dense vector representation for fine-grained retriever. Experiments on the QA dataset show that our model outperforms the most mainstream model greatly by about 11.7% which importantly gets more out of the operation of retriever.
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
References
Chen, D., Fisch, A., Weston, J., Bordes, A.: Reading Wikipedia to answer open-domain questions. arXiv preprint arXiv:1704.00051 (2017)
Chen, T., Kornblith, S., Norouzi, M., Hinton, G.: A simple framework for contrastive learning of visual representations. In: International Conference on Machine Learning, pp. 1597–1607. PMLR (2020)
Clark, C., Gardner, M.: Simple and effective multi-paragraph reading comprehension. arXiv preprint arXiv:1710.10723 (2017)
Das, R., Dhuliawala, S., Zaheer, M., McCallum, A.: Multi-step retriever-reader interaction for scalable open-domain question answering. arXiv preprint arXiv:1905.05733 (2019)
Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: Bert: pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805 (2018)
Gao, T., Yao, X., Chen, D.: SimCSE: simple contrastive learning of sentence embeddings. In: Empirical Methods in Natural Language Processing (EMNLP) (2021)
He, K., Fan, H., Wu, Y., Xie, S., Girshick, R.: Momentum contrast for unsupervised visual representation learning. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 9729–9738 (2020)
Htut, P.M., Bowman, S.R., Cho, K.: Training a ranking function for open-domain question answering. arXiv preprint arXiv:1804.04264 (2018)
Karpukhin, V., et al.: Dense passage retrieval for open-domain question answering. arXiv preprint arXiv:2004.04906 (2020)
Kwiatkowski, T., et al.: Natural questions: a benchmark for question answering research. Trans. Assoc. Comput. Linguist. 7, 453–466 (2019)
Lee, J., Sung, M., Kang, J., Chen, D.: Learning dense representations of phrases at scale. In: Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) (2021)
Lee, J., Sung, M., Kang, J., Chen, D.: Learning dense representations of phrases at scale. arXiv preprint arXiv:2012.12624 (2020)
Lee, J., Wettig, A., Chen, D.: Phrase retrieval learns passage retrieval, too. arXiv preprint arXiv:2109.08133 (2021)
Roberts, A., Raffel, C., Shazeer, N.: How much knowledge can you pack into the parameters of a language model? arXiv preprint arXiv:2002.08910 (2020)
Vaswani, A., et al.: Attention is all you need. In: Advances in Neural Information Processing Systems, vol. 30 (2017)
Voorhees, E.M., et al.: The TREC-8 question answering track report. In: TREC, vol. 99, pp. 77–82 (1999)
Yang, W., et al.: End-to-end open-domain question answering with Bertserini. arXiv preprint arXiv:1902.01718 (2019)
Acknowledgments
We would like to thank the reviewers’s helpful comments. At the same time, we remain gratitude to Jingren Liu for his valuable comments on our work. This work was supported by Projects 61876118 under the National Natural Science Foundation of China, the National Key R &D Program of China under Grant No. 2020AAA0108600 and the Priority Academic Program Development of Jiangsu Higher Education Institutions.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Liu, X., Kong, F. (2022). Coarse-to-Fine Retriever for Better Open-Domain Question Answering. In: Lu, W., Huang, S., Hong, Y., Zhou, X. (eds) Natural Language Processing and Chinese Computing. NLPCC 2022. Lecture Notes in Computer Science(), vol 13551. Springer, Cham. https://doi.org/10.1007/978-3-031-17120-8_31
Download citation
DOI: https://doi.org/10.1007/978-3-031-17120-8_31
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-17119-2
Online ISBN: 978-3-031-17120-8
eBook Packages: Computer ScienceComputer Science (R0)
