通过生成器与判别器协同进化，提升复杂SQL生成的准确率
EvoSQL: Memory-Augmented Critic-Generator Co-Evolution for Text-to-SQL
- 采用记忆增强的生成-判别双模型迭代优化
- 在BIRD数据集上最高提升9.19%准确率
- 适合需要高可靠SQL生成的工业应用
Text-to-SQL 随大语言模型快速发展，但复杂查询仍需多步分解、执行诊断与针对性修正。本文提出EvoSQL，一种生成器与判别器协同进化的框架，通过上下文感知候选记忆库，结合执行反馈与LLM评述验证候选SQL，并基于效用引导聚合更新记忆。为进一步强化生成-判别对，引入自蒸馏策略优化（SDPO）微调阶段，向现代编码类LLM注入执行感知监督。在Spider和BIRD数据集上的实验表明，EvoSQL持续优于Maj@16基线，尤其在BIRD-Dev上提升达+1.37%（Qwen3-4B）至+9.19%（Qwen2.5-Coder-3B）。SDPO初始化进一步提升选定骨干模型在Spider-Test与BIRD-Dev的表现。结果表明，基于记忆的协同进化是实现更可靠、泛化性强Text-to-SQL系统的关键路径。代码已开源。
原文摘要 · Abstract (English)
Text-to-SQL has advanced rapidly with large language models, but complex database queries still require reasoning beyond one-shot generation, including multi-step decomposition, execution-based diagnosis, and targeted correction. We present EvoSQL, a co-evolution framework that formulates SQL synthesis as an iterative interaction between a generator and a critic. EvoSQL maintains a contextualized candidate memory, verifies SQL candidates with both execution signals and LLM-based critique, and updates its memory through utility-guided aggregation. To strengthen the underlying generator-critic pair, we further introduce a Self-Distillation Policy Optimization (SDPO) fine-tuning stage that injects execution-aware supervision into modern coding LLM backbones. Experiments on Spider and BIRD show that EvoSQL consistently improves open-source models over Maj@16 baselines, with particularly large gains on BIRD-Dev, ranging from +1.37% for Qwen3-4B to +9.19% for Qwen2.5-Coder-3B. SDPO initialization further improves selected backbones on Spider-Test and BIRD-Dev. These results suggest that memory-grounded co-evolution is an effective path toward more reliable and generalizable Text-to-SQL systems. Code is available at https://github.com/valleysprings/EvoSQL.
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。