Improving Text-to-SQL with Schema Dependency Learning

TL;DR

This paper introduces SDSQL, a schema dependency learning approach for Text-to-SQL that improves performance and inference speed by reducing reliance on execution-guided decoding, making it more practical for real-world use.

Contribution

The paper proposes a novel schema dependency guided multi-task model that enhances Text-to-SQL performance and inference efficiency without heavily relying on execution-guided decoding.

Findings

Outperforms existing methods on WikiSQL benchmark.

Reduces inference time significantly with minimal performance loss.

Partially replaces the benefits of execution-guided decoding.

Abstract

Text-to-SQL aims to map natural language questions to SQL queries. The sketch-based method combined with execution-guided (EG) decoding strategy has shown a strong performance on the WikiSQL benchmark. However, execution-guided decoding relies on database execution, which significantly slows down the inference process and is hence unsatisfactory for many real-world applications. In this paper, we present the Schema Dependency guided multi-task Text-to-SQL model (SDSQL) to guide the network to effectively capture the interactions between questions and schemas. The proposed model outperforms all existing methods in both the settings with or without EG. We show the schema dependency learning partially cover the benefit from EG and alleviates the need for it. SDSQL without EG significantly reduces time consumption during inference, sacrificing only a small amount of performance and provides more flexibility for downstream applications.