Loading...
Please wait, while we are loading the content...
Similar Documents
Seq2sql: Generating Structured Queries
| Content Provider | Semantic Scholar |
|---|---|
| Author | Zhong, Victor Xiong, Caiming Socher, Richard |
| Copyright Year | 2017 |
| Abstract | Relational databases store a significant amount of the world’s knowledge. However, users are limited in their ability to access this knowledge due to a lack of understanding of query languages such as SQL. We propose Seq2SQL, a deep neural network for translating natural language questions to corresponding SQL queries. Our model leverages the structure of SQL to reduce the output space of generated queries. Moreover, it uses rewards from in-the-loop query execution over the database to learn a policy to generate unordered parts of the query, which are less suitable for optimization via cross entropy loss. In addition, we release WikiSQL, a dataset of 80654 hand-annotated examples of questions and SQL queries distributed across 24241 tables from Wikipedia. This dataset is required to train Seq2SQL and is an order of magnitude larger than comparable datasets. By applying policy-based reinforcement learning with a query execution environment to WikiSQL, Seq2SQL outperforms a state-of-the-art semantic parser by Dong & Lapata (2016), improving execution accuracy from 35.9% to 59.4% and logical form accuracy from 23.4% to 48.3%. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://export.arxiv.org/pdf/1709.00103 |
| Alternate Webpage(s) | https://arxiv.org/pdf/1709.00103v3.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
