arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2512.23366cs.DBcs.AI
收藏

AGRO-SQL:基于高保真数据合成的群体相对优化

AGRO-SQL: Agentic Group-Relative Optimization with High-Fidelity Data Synthesis

Sichuan University(四川大学);IQuest Research(IQuest研究院);Beihang University(北航大学)

Cehua Yang, Dongyu Xiao, Junming Lin, Yuyang Song, Hanxu Yan, Shawn Guo, Wei Zhang, Jian Yang, Mingjie Tang, Bryan Dai

AI总结:

AGRO-SQL通过高保真数据合成和群体相对策略优化,提升文本到SQL系统的推理能力与性能。

摘要:

文本到SQL系统的发展目前受到高质量训练数据稀缺和模型在复杂场景下推理能力有限的阻碍。本文提出一个综合框架,通过双核心方法解决这些问题。从数据角度出发,我们构建了一个迭代数据工厂,合成具有高正确性和精确语义-逻辑对齐的RL准备数据,通过严格验证确保其质量。从模型角度出发,我们引入了新的代理强化学习框架。该框架采用多样性感知的冷启动阶段来初始化稳健的策略,随后通过群体相对策略优化(GRPO)利用环境反馈来完善代理的推理能力。在BIRD和Spider基准测试中,我们的协同方法在单模型方法中实现了最先进的性能。

英文摘要:

The advancement of Text-to-SQL systems is currently hindered by the scarcity of high-quality training data and the limited reasoning capabilities of models in complex scenarios. In this paper, we propose a holistic framework that addresses these issues through a dual-centric approach. From a Data-Centric perspective, we construct an iterative data factory that synthesizes RL-ready data characterized by high correctness and precise semantic-logic alignment, ensured by strict verification. From a Model-Centric perspective, we introduce a novel Agentic Reinforcement Learning framework. This framework employs a Diversity-Aware Cold Start stage to initialize a robust policy, followed by Group Relative Policy Optimization (GRPO) to refine the agent's reasoning via environmental feedback. Extensive experiments on BIRD and Spider benchmarks demonstrate that our synergistic approach achieves state-of-the-art performance among single-model methods.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑