An Efficient and Effective Evaluator for Text2SQL Models on Unseen and Unlabeled Data

TL;DR

FusionSQL is a novel evaluation method that estimates Text2SQL system accuracy on unseen, unlabeled data by analyzing output patterns, enabling early detection of quality issues without requiring labeled datasets.

Contribution

It introduces a model-agnostic approach to assess Text2SQL performance on unlabeled data, addressing a key deployment challenge.

Findings

FusionSQL accurately tracks actual system accuracy.

It reliably detects quality decline in real-world scenarios.

Supports continuous monitoring and pre-release checks.

Abstract

Recent advances in large language models has strengthened Text2SQL systems that translate natural language questions into database queries. A persistent deployment challenge is to assess a newly trained Text2SQL system on an unseen and unlabeled dataset when no verified answers are available. This situation arises frequently because database content and structure evolve, privacy policies slow manual review, and carefully written SQL labels are costly and time-consuming. Without timely evaluation, organizations cannot approve releases or detect failures early. FusionSQL addresses this gap by working with any Text2SQL models and estimating accuracy without reference labels, allowing teams to measure quality on unseen and unlabeled datasets. It analyzes patterns in the system's own outputs to characterize how the target dataset differs from the material used during training. FusionSQL supports pre-release checks, continuous monitoring of new databases, and detection of quality decline. Experiments across diverse application settings and question types show that FusionSQL closely follows actual accuracy and reliably signals emerging issues. Our code is available at https://github.com/phkhanhtrinh23/FusionSQL.