arXiv：cs.CL· Wen-Zhi Li, Yue Gong, Konstantinos Kanellis, Balakrishnan Murali Narayanaswamy·· 4 小时前AI 评分39
超越正确性：解决 Agentic Text-to-SQL 中的欠规范问题
Beyond Correctness: Resolving Underspecification in Agentic Text-to-SQL
AI 导读
研究指出 Agentic Text-to-SQL 系统即使执行结果正确，也可能因过早终止澄清而默默做出未经验证的假设。为此提出 PlanPool，将澄清计划外化为可变问题池，每个计划问题必须被明确提问或丢弃，交互中发现的新歧义可随时加入。在 BIRD-Interact 和 Spider 衍生的三个基准上，PlanPool 持续提升歧义覆盖率并减少静默失败，同时保持有竞争力的执行准确率。
正文
Abstract:Agentic Text-to-SQL systems can interact with users to clarify underspecified queries before generating SQL. However, a correct execution result does not necessarily imply that the agent has adequately resolved the underlying underspecification: the agent may silently make unverified assumptions that happen to match the intended answer. We show that this behavior is driven in part by premature clarification termination. Although forcing an agent to ask more questions improves execution accuracy, ambiguities are concentrated in earlier interactions, making brute-force questioning inefficient. More importantly, even when explicitly prompted to plan its clarification process, the agent frequently abandons questions that it has already identified as relevant. To address this failure mode, we introduce PlanPool, which externalizes the clarification plan as a mutable question pool. Every planned question must be explicitly asked or dropped before submission, while newly discovered ambiguities can be added during interaction. Across three benchmarks derived from BIRD-Interact and Spider, PlanPool consistently improves ambiguity coverage and reduces silent failures over unconstrained and prompt-based alternatives, while maintaining competitive execution accuracy. Our results highlight an important distinction in agentic reasoning: identifying missing information is not sufficient, and the agent must also reliably maintain and resolve it before committing to an answer.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.02739 [cs.CL] |
| (or arXiv:2610.02739v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.02739 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Wen-Zhi Li [view email]
[v1]
Fri, 2 Oct 2026 03:10:55 UTC (181 KB)
来源：arXiv：cs.CL · arxiv.org