arXiv：cs.LG· Arijit Sehanobish, Bruno Gomes Coelho, Guillaume Michel, Sophia Zhi, Valerie Faucon-Morin, Kristen Howell·· 5 小时前AI 评分40
FLINT：面向复杂金融数据的生产级 Text-to-SQL 系统
From Benchmarks to Production: A Text-to-SQL System for Complex Financial Data
AI 导读
针对生产金融数据库中概念以不透明整数键存储、通用 Text-to-SQL 系统准确率不足 50% 的问题，研究者提出领域专用系统 FLINT，通过查找智能体、查询模板嵌入检索和外键链模式链接三个组件缩小差距。在总计 359 个问题的两个生产金融模式数据集上，FLINT 使用同一 LLM 即超越多种 SOTA 基线，并已作为金融数据检索服务的一部分部署在生产环境。
正文
Abstract:General-purpose Text-to-SQL systems achieve strong performance on academic benchmarks like Spider and BIRD, where schemas are relatively shallow and column values are often human readable. In production financial databases, where concepts are stored as opaque integer keys rather than human-readable strings, these methods fall below 50%, as even simple queries require multiple joins and filter predicates reference opaque IDs. We present Financial LINking Text-to-SQL (FLINT), a domain-specialized Text-to-SQL system that closes this gap through three key components: (1) a lookup agent that dynamically resolves natural-language concepts to question-specific reference table constraints, (2) embedding-based retrieval of structurally similar query templates from a compact, expert-authored bank, and (3) schema linking that prunes a large table schema to the relevant subset by traversing foreign-key chains, rather than relying on name similarity alone. We evaluate on two datasets totaling 359 questions over production financial schemas. FLINT outperforms various state-of-the-art baselines using the same LLM. The system is deployed in production as part of a financial data retrieval service.
| Comments: | EMNLP Industry Track 2026 |
| Subjects: | Artificial Intelligence (cs.AI); Databases (cs.DB); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.03524 [cs.AI] |
| (or arXiv:2610.03524v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.03524 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Arijit Sehanobish [view email]
[v1]
Fri, 2 Oct 2026 16:13:40 UTC (91 KB)
来源：arXiv：cs.LG · arxiv.org