作者位次 #9/12ACL 2026Poster· Natural Language Processing
SQL -Trail: Multi-Turn Reinforcement Learning with Interleaved Feedback for Text-to- SQL
Harper Hua · Zhen Han · Zhengyuan Shen · Meng-Chieh Lee · Sheng Guan · Qi Zhu · Sullam Jeoung · Yueyan Chen · Yunfei Bai · Shuai Wang · Vassilis N. Ioannidis · Huzefa Rangwala
摘要
While large language models (LLMs) have substantially improved Text-to-SQL generation, a pronounced gap remains between AI systems and human experts on challenging benchmarks such as BIRD-SQL. We argue this gap stems largely from the prevailing single-pass paradigm, which lacks the iterative reas...