Progress-SQL: Improving Reinforcement Learning for Text-to-SQL via Progressive Rewards
Quick Answer
Progress-SQL enhances Text-to-SQL generation by introducing a multi-turn reinforcement learning framework with progressive rewards, improving performance on benchmarks like BIRD and Spider.
Quick Take
Progress-SQL enhances Text-to-SQL generation by introducing a multi-turn reinforcement learning framework with progressive rewards, improving performance on benchmarks like BIRD and Spider. The method employs an Oracle-guided Diagnostic Tree for structured feedback and combines multiple reward signals, leading to consistent performance gains in both primary and robustness evaluations.
Key Points
- Introduces Oracle-guided Diagnostic Tree for clause-level SQL abstraction.
- Combines structural and lexical alignment for dense reward signals.
- Implements progressive rewards to measure SQL improvement iteratively.
- Demonstrates consistent performance gains on BIRD and Spider benchmarks.
- Incorporates latency and execution status rewards for enhanced SQL correction.
Article Content
From source RSS / original summaryarXiv:2606. 06825v1 Announce Type: new Abstract: Reinforcement learning has recently shown promise in improving large language models for Text-to-SQL generation, yet existing methods typically optimize one-shot rewards defined over a single SQL state. Such rewards provide limited guidance for iterative SQL correction and are insufficient to capture the improvement of multi-turn SQL refinement.
In this paper, we propose Progress-SQL, a multi-turn reinforcement learning framework with progressive rewards for Text-to-SQL. Our approach introduces an Oracle-guided Diagnostic Tree (ODT), which abstracts SQL queries into clause-level structural profiles and produces diagnostic feedback for next-turn refinement.
To provide dense and robust reward signals, we combine ODT-based structural alignment with lexical alignment and define a progressive reward that measures the improvement from the initial SQL to the final SQL. We further incorporate a progression latency reward that favors earlier correctness and an execution status reward that encourages recovery from the invalid SQL.
Experiments on BIRD, Spider, and Spider robustness variants demonstrate that our method consistently improves Text-to-SQL performance across both primary and robustness evaluations.
Reader Mode unavailable (could not extract clean content).
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →Time to REFLECT: Can We Trust LLM Judges for Evidence-based Research Agents?
The REFLECT benchmark reveals that current LLM judges are unreliable, achieving below 55% accuracy in evaluating reasoning and evidence use, highlighting the need for improved evaluation methods for deep research agents.