Towards Complex Text-to-SQL in Cross-Domain Database with Intermediate Representation

TL;DR

IRNet introduces a three-phase neural approach for complex cross-domain Text-to-SQL translation, using an intermediate representation to improve accuracy and interpretability, achieving state-of-the-art results on the Spider benchmark.

Contribution

The paper proposes IRNet, a novel multi-phase model with an intermediate SemQL representation, enhancing cross-domain Text-to-SQL performance over previous methods.

Findings

IRNet achieves 46.7% accuracy on Spider benchmark.

IRNet outperforms previous state-of-the-art by 19.5%.

IRNet attains first position on the Spider leaderboard.

Abstract

We present a neural approach called IRNet for complex and cross-domain Text-to-SQL. IRNet aims to address two challenges: 1) the mismatch between intents expressed in natural language (NL) and the implementation details in SQL; 2) the challenge in predicting columns caused by the large number of out-of-domain words. Instead of end-to-end synthesizing a SQL query, IRNet decomposes the synthesis process into three phases. In the first phase, IRNet performs a schema linking over a question and a database schema. Then, IRNet adopts a grammar-based neural model to synthesize a SemQL query which is an intermediate representation that we design to bridge NL and SQL. Finally, IRNet deterministically infers a SQL query from the synthesized SemQL query with domain knowledge. On the challenging Text-to-SQL benchmark Spider, IRNet achieves 46.7% accuracy, obtaining 19.5% absolute improvement over previous state-of-the-art approaches. At the time of writing, IRNet achieves the first position on the Spider leaderboard.