Boundary-Aware NL2SQL: Integrating Reliability through Hybrid Reward and Data Synthesis

2026-01-15

AI Summary

This work addresses the critical limitation of natural language to SQL (NL2SQL) systems in reliably rejecting out-of-domain or semantically ambiguous queries due to insufficient boundary awareness. To this end, we propose BAR-SQL, a novel framework that constructs enterprise-scale corpora enriched with boundary cases through seed mutation synthesis, and integrates knowledge-anchored reasoning to generate interpretable chains of thought. BAR-SQL employs a two-stage training strategy—supervised fine-tuning followed by Grouped Relative Policy Optimization (GRPO)—augmented with a task-conditioned hybrid reward mechanism that jointly optimizes SQL execution accuracy and rejection semantic precision. Our approach pioneers the explicit incorporation of boundary awareness into the NL2SQL generation pipeline and introduces Ent-SQL-Bench, the first benchmark evaluating both generation and rejection capabilities. On this benchmark, BAR-SQL achieves a state-of-the-art average accuracy of 91.48%, outperforming advanced closed-source models such as Claude 4.5 Sonnet and GPT-5.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Search and Optimization: Learning to Search
• Knowledge Representation and Reasoning: Common-Sense Reasoning

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

In this paper, we present BAR-SQL (Boundary-Aware Reliable NL2SQL), a unified training framework that embeds reliability and boundary awareness directly into the generation process. We introduce a Seed Mutation data synthesis paradigm that constructs a representative enterprise corpus, explicitly encompassing multi-step analytical queries alongside boundary cases including ambiguity and schema limitations. To ensure interpretability, we employ Knowledge-Grounded Reasoning Synthesis, which produces Chain-of-Thought traces explicitly anchored in schema metadata and business rules. The model is trained through a two-stage process: Supervised Fine-Tuning (SFT) followed by Reinforcement Learning via Group Relative Policy Optimization. We design a Task-Conditioned Hybrid Reward mechanism that simultaneously optimizes SQL execution accuracy-leveraging Abstract Syntax Tree analysis and dense result matching-and semantic precision in abstention responses. To evaluate reliability alongside generation accuracy, we construct and release Ent-SQL-Bench, which jointly assesse SQL precision and boundary-aware abstention across ambiguous and unanswerable queries. Experimental results on this benchmark demonstrate that BAR-SQL achieves 91.48% average accuracy, outperforming leading proprietary models, including Claude 4.5 Sonnet and GPT-5, in both SQL generation quality and boundary-aware abstention capability. The source code and benchmark are available anonymously at: https://github.com/TianSongS/BAR-SQL.

Problem

• NL2SQL
• boundary awareness
• reliability
• abstention
• ambiguous queries

Innovation

• Boundary-Aware NL2SQL
• Hybrid Reward
• Data Synthesis
• Knowledge-Grounded Reasoning
• Reliable Abstention