arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2608.22651cs.CL

无需复杂设计的迭代:简单的ReAct架构足以完成文本转SQL生成

Iteration Without Elaboration: A Simple ReAct Architecture Suffices for Text-to-SQL Generation

  • Duke University(杜克大学)

机构由 AI 辅助整理,请以论文原文为准。

Jian Lu, Haiwei Yu, Raymond M Xiong, Anru Zhang, Danyang Zhuo

AI总结:

该研究针对现有文本转SQL系统复杂且延迟高的问题,提出简单的ReAct-SQL框架,在两个数据集上达到高准确率且运行速度提升8倍,验证了其有效性。

AI中文摘要:

现代文本转SQL系统变得日益复杂,依赖模式链接模块、检索增强提示、候选生成以及多阶段优化流水线,虽有效却带来了显著的延迟和工程开销。为此,我们提出ReAct-SQL,这是一种简单却有效的零样本ReAct风格框架,仅基于迭代推理和由15种关系操作组成的类型化领域特定语言(DSL)定义的受限动作空间,而非自由形式的SQL生成。模型逐步发出DSL调用,观察编译后SQL的执行反馈,并通过交互修正推理。在修正后的BIRD mini-dev和EHR-SQL数据集上,ReAct-SQL分别达到84.5%和73.9%的准确率,与复杂得多的基线模型表现相当,同时运行速度最高快8倍。增量消融实验进一步表明,迭代主要提升了 grounding( grounding可译为“ grounding:指模型将语言表述与实际数据或操作对应匹配的能力”),而DSL则提升了组合可靠性。

英文摘要:

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.

补充信息

↑