X-SQL: reinforce schema representation with context

TL;DR

Addresses NL-to-SQL semantic parsing on WikiSQL. Proposes X-SQL, a three-layer architecture that enriches schema representations with BERT-style contextual outputs and type embeddings, and uses a KL-divergence objective for WHERE predictions. Shows state-of-the-art results on WikiSQL with and without execution guidance, including surpassing 90% logical-form accuracy on the test set. Demonstrates that contextualized schema representations improve downstream SQL generation and suggests applicability to more complex datasets.

Abstract

In this work, we present X-SQL, a new network architecture for the problem of parsing natural language to SQL query. X-SQL proposes to enhance the structural schema representation with the contextual output from BERT-style pre-training model, and together with type information to learn a new schema representation for down-stream tasks. We evaluated X-SQL on the WikiSQL dataset and show its new state-of-the-art performance.

Figures (1)

• Figure 1: Components of X-SQL neural network model architecture.