Overview / Usage

A significant amount of the world's knowledge is stored in relational databases. However, the ability for users to retrieve facts from a database is limited due to a lack of understanding of query languages such as SQL. I am trying to study and propose extended models of Seq2SQL, In2SQL and text to sql ,trying to modify using different approaches.A deep neural network for translating natural language questions to corresponding SQL queries. Our model leverages the structure of SQL queries to significantly reduce the output space of generated queries. Moreover, we use rewards from in-the-loop query execution similar to seq to sql model over the database to learn a policy to generate unordered parts of the query, which we show are less suitable for optimization via cross entropy loss. In addition, You can search for published WikiSQL, a dataset of 80654 hand-annotated examples of questions and SQL queries distributed across 24241 tables from Wikipedia.

If anybody is interested for same can contact me.

Methodology / Approach

The existing approach is to generate the query from the knowledge of SQL manually. But certain improvement done in recent years helps to generate more accurate queries using Probabilistic Context Free Grammar (PCFG). The current implemented standard is QuePy [10] and similar, disjoint projects like them.

Although in meantime tech giants Sales force and Microsoft have tried to work on similar topics to extent and they have successfully been able to publish research papers which are sited below. I would like to take similar approach and study there procedures and will try to implement them using Intel technologies like Movidius stick and devcloud. Usage of Gans and Reinforcement learning are hoped to provide beneficial results. If possible I would like to extend project to construction of complex qureries too.

Papers:

• Seq2SQL: Generating Structured Queries from Natural Language using Reinforcement Learningsalesforce
• Neural Enquirer: Learning to Query Tables in Natural Language

Dataset Example:

A large annotated semantic parsing corpus for developing natural language interfaces.

​

​

Technologies Used

Python 3+ and all basic libraries

CNN , RNN ,GANS

Devcloud,Movidius stick

​

Other links

• Natural Language text to SQL query
• Seq2SQL: Generating Structured Queries from Natural Language using Reinforcement Learning