الملخص
State-of-the-art Text-to-SQL systems are typically multi-agent pipelines centered around two fundamental tasks: schema linking and SQL generation. However, existing work trains separate models for each task, failing to leverage the synergy between these interrelated tasks. In this work, we propose DualSQL, a new Text-to-SQL system consisting of two agents powered by a single model backbone. The agents share the same model weights and agentic scaffold, enabling joint optimization through a robust multi-agent reinforcement learning (RL) framework. We design three database access tools to facilitate effective multi-step reasoning grounded to interactions with the databases. To improve training and avoid model collapse, we introduce a set of rollout guardrail mechanisms that stabilizes multi-agent RL training, supporting DualSQL to keep improving during training. We also introduce a new SQL correctness metric, robust execution match (REX), to more accurately judge SQL correctness and assign reward signals. Being trained on only 3755 examples, DualSQL-4B achieves an impressive 68.0% execution accuracy on the BIRD development set, matching previous 7B models. DualSQL-8B further improves to 71.1%, outperforming previous state-of-the-art single-model solutions with 32B parameters. These results demonstrate the strength of joint multi-agent reinforcement learning for building high performance Text-to-SQL pipelines.
الكلمات المفتاحية
الموضوع
بيانات النشر
- المجلة
- غير متاح
- وصول مفتوح
- وصول مفتوح أخضر
اقتبس هذه المقالة
APA 7
Chen, S., Gan, Y., Chung, Y., Zhang, J., Li, Q., Bodapati, S. B., Greer, C. J., Su, Y., & Ozcan, F. (2026). DualSQL: Text-to-SQL with Multi-Agent Reinforcement Learning. https://omanscience.com/ar/articles/dualsql-text-to-sql-with-multi-agent-reinforcement-learning
MLA 9
Chen, Shijie, et al. "DualSQL: Text-to-SQL with Multi-Agent Reinforcement Learning." https://omanscience.com/ar/articles/dualsql-text-to-sql-with-multi-agent-reinforcement-learning.
شيكاغو (المؤلف–التاريخ)
Chen, Shijie, Yu Gan, Yeounoh Chung, Jiani Zhang, Quannan Li, Sravan Babu Bodapati, Cody J. Greer, Yu Su, and Fatma Ozcan. 2026. "DualSQL: Text-to-SQL with Multi-Agent Reinforcement Learning." https://omanscience.com/ar/articles/dualsql-text-to-sql-with-multi-agent-reinforcement-learning.
هارفارد
Chen, S., Gan, Y., Chung, Y., Zhang, J., Li, Q., Bodapati, S. B., Greer, C. J., Su, Y. and Ozcan, F. (2026) 'DualSQL: Text-to-SQL with Multi-Agent Reinforcement Learning', Available at: https://omanscience.com/ar/articles/dualsql-text-to-sql-with-multi-agent-reinforcement-learning.
فانكوفر
Chen S, Gan Y, Chung Y, Zhang J, Li Q, Bodapati SB, et al. DualSQL: Text-to-SQL with Multi-Agent Reinforcement Learning. https://omanscience.com/ar/articles/dualsql-text-to-sql-with-multi-agent-reinforcement-learning
IEEE
S. Chen, Y. Gan, Y. Chung, J. Zhang, Q. Li, S. B. Bodapati, C. J. Greer, Y. Su, and F. Ozcan, "DualSQL: Text-to-SQL with Multi-Agent Reinforcement Learning," https://omanscience.com/ar/articles/dualsql-text-to-sql-with-multi-agent-reinforcement-learning.