用简单迭代框架实现高效文本转SQL，速度提升8倍且精度不降。
Iteration Without Elaboration: A Simple ReAct Architecture Suffices for Text-to-SQL Generation
- 基于15种关系操作的约束语言，通过迭代推理逐步生成SQL。
- 在BIRD和EHR-SQL上分别达到84.5%和73.9%准确率，超越复杂模型。
- 适合追求低延迟、高可靠性的实际部署场景。
当前文本转SQL系统日趋复杂，依赖模式链接模块、检索增强提示、候选生成与多阶段精炼流程。这些设计虽有效，但带来显著延迟与工程开销。为此，我们提出ReAct-SQL，一种仅基于迭代推理与15种关系操作构成的受限领域特定语言（DSL）的零样本框架。模型逐步调用DSL指令，观察编译后SQL的执行反馈，并据此修正推理。在修正后的BIRD mini-dev和EHR-SQL数据集上，ReAct-SQL分别取得84.5%和73.9%的准确率，性能媲美更复杂的基线，同时运行速度最快达8倍提升。增量消融实验表明，迭代主要增强任务对齐性，而DSL则提升组合可靠性。
原文摘要 · Abstract (English)
Modern text-to-SQL systems have become increasingly elaborate, relying on schema-linking modules, retrieval-augmented prompting, candidate generation, and multi-stage refinement pipelines. While effective, these additions introduce substantial latency and engineering overhead. To this end, we present \textbf{ReAct-SQL}, a simple yet effective zero-shot ReAct-style framework built solely on iterative reasoning and a constrained action space defined by a typed Domain-Specific Language (DSL) of 15 relational operations, rather than free-form SQL generation. The model incrementally issues DSL calls, observes compiled-SQL execution feedback, and revises its reasoning through interaction. On corrected BIRD mini-dev and EHR-SQL, ReAct-SQL achieves \textbf{84.5\%} and \textbf{73.9\%} accuracy, respectively, matching substantially more elaborate baselines while running up to $8\times$ faster. Incremental ablations further show that iteration primarily improves grounding, while the DSL improves compositional reliability.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。