EHR-SeqSQL : A Sequential Text-to-SQL Dataset For Interactively Exploring Electronic Health Records

Anonymous

Abstract: Text-to-SQL parsing allows non-expert users to obtain information from databases using natural language. There are several important yet under-explored objectives in this field: interactivity, compositionality, and efficiency. In this paper, we present EHR-SeqSQL, a sequential text-to-SQL dataset, specifically designed for Electronic Health Record (EHR) databases. We demonstrate the benefits of multi-turn setting over single-turn setting with respect to compositionality, and provide a new data split and an additional test set to evaluate compositional generalization. Furthermore, we introduce unique special tokens in SQL queries to enhance execution efficiency. By addressing all these objectives above, our research aims to bridge the gap between industry needs and academic research in the text-to-SQL domain.

Paper Type: long

Research Area: Dialogue and Interactive Systems

Contribution Types: Data resources, Data analysis

Languages Studied: English

Consent To Share Submission Details: On behalf of all authors, we agree to the terms above to share our submission details.