基于问句语义图神经网络的中文问句生成SQL语句研究
Research on Chinese Question Generation SQL Statement Based on Question Semantic Graph Neural Network
摘要: 自然语言问句转为结构化查询语句(Text-to-SQL)是语义解析领域中热点研究之一，其目标是将自然语言问句转化为数据库可以理解且执行的结构化查询语句。现有研究大部分仅考虑数据库层面的关联信息，忽略了问句中的实体关系信息的重要性。为了提高模型捕捉问句中语义的有用信息，本文在IGSQL模型基础上，引入问句中实体之间的图网络信息，通过注意力机制来自动学习问句和数据库模式之间的关联。在Chase数据集上的实验结果表明，本文提出模型的完全匹配率达到46.2%。相比较于基线模型，完全匹配率提升了6.3%。
Abstract: The conversion of natural language questions into structured query statements (Text to SQL) is one of the hot research topics in the field of semantic parsing, with the goal of transforming natural language questions into structured query statements that can be understood and executed by databases. Most existing research only considers relational information at the database level, ignoring the importance of entity relationship information in questions. In order to improve the model’s ability to capture useful semantic information in questions, this paper introduces graph network information between entities in questions based on the IGSQL model, and automatically learns the association between questions and database patterns through attention mechanisms. The experimental results on the Chase dataset show that the proposed model has an exact matching accuracy of 46.2%. Compared to the baseline model, the exact matching accuracy has increased by 6.3%.
文章引用：张海芳, 何清龙. 基于问句语义图神经网络的中文问句生成SQL语句研究[J]. 运筹与模糊学, 2024, 14(1): 83-90. https://doi.org/10.12677/ORF.2024.141008

参考文献

[1] 潘璇, 徐思涵, 蔡祥睿, 等. 基于深度学习的数据库自然语言接口综述[J]. 计算机研究与发展, 2021, 58(9): 1925-1950.
[2] Yang, J., Jiang, H., Yin, Q., et al. (2022) Seqzero: Few-Shot Compositional Semantic Parsing with Sequential Prompts and Zero-Shot Models. Findings of the Association for Computational Linguistics, Seattle, July 2022, 49-60. [Google Scholar] [CrossRef] 
[3] Iyer, S., Cheung, A. and Zettlemoyer, L. (2019) Learning Programmatic Idioms for Scalable Semantic Parsing. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Pro-cessing, Hong Kong, November 2019, 5426-5435. [Google Scholar] [CrossRef] 
[4] Li, D. and Lapata, M. (2016) Language to Logical Form with Neural Attention. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, Volume 1, 33-43.
[5] 郑耀东, 李旭峰, 陈和平, 贺桂娇. 基于中文自然语言的SQL生成综述[J]. 计算机系统应用, 2023, 32(12): 32-42.
[6] Wu, Z., Pan, S., Chen, F., et al. (2021) A Comprehensive Survey on Graph Neural Networks. IEEE Transactions on Neural Networks and Learning Systems, 32, 4-24. [Google Scholar] [CrossRef] 
[7] Xu, X., Liu, C. and Song, D. (2017) SQLNet: Generating Structured Queries from Natural Language without Reinforcement Learning.
[8] Bogin, B., Gardner, M. and Berant, J. (2019) Representing Schema Structure with Graph Neural Networks for Text-to-SQL Parsing. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, Florence, July 2019, 4560-4565. [Google Scholar] [CrossRef] 
[9] Bogin, B., Gardner, M. and Berant, J. (2019) Global Reasoning over Database Structures for Text-to-SQL Parsing. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), Hong Kong, November 2019, 3659-3664. [Google Scholar] [CrossRef] 
[10] Zhong, V., Xiong, C. and Socher, R. (2017) Seq2SQL: Generating Structured Queries from Natural Language Using Reinforcement Learning.
[11] Dong, L. and Lapata, M. (2016) Language to Logical Form with Neural Attention. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, Volume 1, 33-43. [Google Scholar] [CrossRef] 
[12] Yu, T., Yasunaga, M., Yang, K., et al. (2018) SyntaxSQLNet: Syntax Tree Networks for Complex and Cross-Domain Text-to-SQL Task. Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, Brussels, Octo-ber-November 2018, 1653-1663. [Google Scholar] [CrossRef] 
[13] 邱锡鹏. 神经网络与深度学习[M]. 北京: 机械工业出版社, 2020.
[14] Che, W., Feng, Y., Qin, L., et al. (2021) N-LTP: An Open-Source Neural Language Technology Platform for Chinese. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, November 2021, 42-49. [Google Scholar] [CrossRef] 
[15] Liu, Y., Ott, M., Goyal, N., et al. (2019) RoBERTa: A Robustly Optimized BERT Pretraining Approach.
[16] Paszke, A., Gross, S., Massa, F., et al. (2019) PyTorch: An Imperative Style, High-Performance Deep Learning Library. 33rd Annual Conference on Neural Information Pro-cessing Systems (NeurIPS 2019), Vancouver, 8-14 December 2019, 8026-8037.
[17] Guo, J.Q., Si, Z.L., Wang, Y., et al. (2021) Chase: A Large-Scale and Pragmatic Chinese Dataset for Cross-Database Context-Dependent Text-to-SQL. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, Volume 1, 2316-2331. [Google Scholar] [CrossRef] 
[18] 赵志超, 游进国, 何培蕾, 李晓武. 数据库中文查询对偶学习式生成SQL语句研究[J]. 中文信息学报, 2023, 37(3): 164-172.
[19] Zhang, R., Yu, T., et al. (2019) Edit-ing-Based SQL Query Generation for Cross-Domain Context-Dependent Questions. Proceedings of the 2019 Conference on EMNLP and the 9th IJCNLP, Hong Kong, November 2019, 5338-5349. [Google Scholar] [CrossRef] 
[20] Cai, Y.T. and Wan, X.J. (2020) IGSQL: Database Schema Inter-action Graph Based Neural Model for Context-Dependent Text-to-SQL Generation. Proceedings of the 2020 Con-ference on EMNLP, November 2020, 6903- 6912. [Google Scholar] [CrossRef] 

为你推荐