Tracking Interaction States for Multi-Turn Text-to-SQL Semantic Parsing

TL;DR

This paper introduces a novel approach for multi-turn text-to-SQL semantic parsing that explicitly tracks interaction states based on schema items and SQL keywords, improving decoding accuracy.

Contribution

It defines and models interaction states with a relational graph neural network, enhancing the translation of natural language to SQL in multi-turn dialogues.

Findings

Outperforms existing methods on the CoSQL dataset

Achieves state-of-the-art results on the task leaderboard

Demonstrates the effectiveness of explicit interaction state modeling

Abstract

The task of multi-turn text-to-SQL semantic parsing aims to translate natural language utterances in an interaction into SQL queries in order to answer them using a database which normally contains multiple table schemas. Previous studies on this task usually utilized contextual information to enrich utterance representations and to further influence the decoding process. While they ignored to describe and track the interaction states which are determined by history SQL queries and are related with the intent of current utterance. In this paper, two kinds of interaction states are defined based on schema items and SQL keywords separately. A relational graph neural network and a non-linear layer are designed to update the representations of these two states respectively. The dynamic schema-state and SQL-state representations are then utilized to decode the SQL query corresponding to current utterance. Experimental results on the challenging CoSQL dataset demonstrate the effectiveness of our proposed method, which achieves better performance than other published methods on the task leaderboard.