LG AI Research & KAIST at EHRSQL 2024: Self-Training Large Language Models with Pseudo-Labeled Unanswerable Questions for a Reliable Text-to-SQL System on EHRs

TL;DR

This paper introduces a self-training approach with pseudo-labeled unanswerable questions to improve the reliability of text-to-SQL models for Electronic Health Records, emphasizing accurate identification of unanswerable queries.

Contribution

It proposes a novel two-stage training and filtering method using pseudo-labeled unanswerable questions to enhance model reliability in healthcare text-to-SQL tasks.

Findings

Achieved top performance in EHRSQL 2024 shared task.

Improved model accuracy in identifying unanswerable questions.

Enhanced reliability of text-to-SQL systems for EHRs.

Abstract

Text-to-SQL models are pivotal for making Electronic Health Records (EHRs) accessible to healthcare professionals without SQL knowledge. With the advancements in large language models, these systems have become more adept at translating complex questions into SQL queries. Nonetheless, the critical need for reliability in healthcare necessitates these models to accurately identify unanswerable questions or uncertain predictions, preventing misinformation. To address this problem, we present a self-training strategy using pseudo-labeled unanswerable questions to enhance the reliability of text-to-SQL models for EHRs. This approach includes a two-stage training process followed by a filtering method based on the token entropy and query execution. Our methodology's effectiveness is validated by our top performance in the EHRSQL 2024 shared task, showcasing the potential to improve healthcare decision-making through more reliable text-to-SQL systems.