Translating Natural Language to SQL using Pointer-Generator Networks and How Decoding Order Matters

  • 2018-11-13 14:06:58
  • Denis Lukovnikov, Nilesh Chakraborty, Jens Lehmann, Asja Fischer
  • 1

Abstract

Translating natural language to SQL queries for table-based questionanswering is a challenging problem and has received significant attention fromthe research community. In this work, we extend a pointer-generator andinvestigate the order-matters problem in semantic parsing for SQL. Even thoughour model is a straightforward extension of a general-purposepointer-generator, it outperforms early works for WikiSQL and remainscompetitive to concurrently introduced, more complex models. Moreover, weprovide a deeper investigation of the potential order-matters problem thatcould arise due to having multiple correct decoding paths, and investigate theuse of REINFORCE as well as a dynamic oracle in this context.

 

Quick Read (beta)

loading the full paper ...