Synthetic SQL Mastery Bridging Data Quality and Complex Reasoning Gaps
📝 Original Paper Info
- Title: AGRO-SQL Agentic Group-Relative Optimization with High-Fidelity Data Synthesis- ArXiv ID: 2512.23366
- Date: 2025-12-29
- Authors: Cehua Yang, Dongyu Xiao, Junming Lin, Yuyang Song, Hanxu Yan, Shawn Guo, Wei Zhang, Jian Yang, Mingjie Tang, Bryan Dai
📝 Abstract
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.💡 Summary & Analysis
1. [[Key Contribution 1 in English]] - [[Simple Explanation with Metaphor in English]]: [[Sci-Tube Style Script in English]] 2. [[Key Contribution 2 in English]] - [[Simple Explanation with Metaphor in English]]: [[Sci-Tube Style Script in English]] 3. [[Key Contribution 3 in English]] - [[Simple Explanation with Metaphor in English]]: [[Sci-Tube Style Script in English]]📄 Full Paper Content (ArXiv Source)
📊 논문 시각자료 (Figures)
