An Encoder-Decoder Framework Translating Natural Language to Database Queries

Authors: Ruichu Cai, Boyan Xu, Zhenjie Zhang, Xiaoyan Yang, Zijian Li, Zhihao Liang

IJCAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The empirical evaluation on real world database and queries show that our approach outperform state-of-the-art solution by a significant margin.
Researcher Affiliation Collaboration 1 Faculty of Computer, Guangdong University of Technology, China 2 Singapore R&D, Yitu Technology Ltd.
Pseudocode No The paper describes the techniques in detail but does not include any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any explicit statement about making the source code available or include a link to a code repository.
Open Datasets Yes We run our experiments on three databases, namely Geo880, Academic and IMDB. Academic database has 17 tables, collected by Microsoft Academic Search [Roy et al., 2013]. We reuse the queries and natural language descriptions in Geo880 database, and use the standard training/test split as in [Iyer et al., 2017].
Dataset Splits Yes All numbers reported in the experiments on MAS and IMDB are average of 5-fold cross validations. We reuse the queries and natural language descriptions in Geo880 database, and use the standard training/test split as in [Iyer et al., 2017].
Hardware Specification No The paper mentions training models but does not specify any details about the hardware used (e.g., specific GPU/CPU models, memory amounts).
Software Dependencies Yes Our model is implemented in TensorFlow 1.2.0.
Experiment Setup Yes We optimize the hyerparameters in all approaches and use the configuration with best results. The result hyperparameters are listed in Table 3.