SemanticAgent: A Semantics-Aware Framework for Text-to-SQL Data Synthesis
For text-to-SQL researchers and practitioners, SemanticAgent improves the semantic quality of synthetic training data, addressing a known bottleneck in data synthesis.
SemanticAgent addresses the problem that existing text-to-SQL synthesis pipelines conflate executability with semantic validity, retaining queries that execute but violate database semantics. The framework generates synthetic data that outperforms prior methods under semantic-quality evaluation, leading to stronger downstream fine-tuning performance on semantically demanding benchmarks.
Existing text-to-SQL synthesis pipelines still conflate executability with semantic validity: syntactic checks and execution-based validation can retain queries that execute successfully while violating database semantics. To address these limitations, we propose SemanticAgent, a semantic-aware synthesis framework. SemanticAgent organizes synthesis around three specialized modules: an analyzer, a synthesizer, and a verifier. Through a three-stage protocol of semantic analysis, stepwise synthesis, and diagnostic refinement, SemanticAgent transforms execution-based validation alone into a traceable reasoning process. Our framework generates synthetic data that consistently outperforms prior synthesis methods under semantic-quality evaluation, leading to stronger downstream fine-tuning performance, especially on semantically demanding benchmarks.