DCG-SQL: Enhancing In-Context Learning for Text-to-SQL with Deep Contextual Schema Link Graph
Jihyung Lee, Jin-Seop Lee, Jaehoon Lee, YunSeok Choi, Jee-Hyong Lee · 2025
Read DCG-SQL: Enhancing In-Context Learning for Text-to-SQL with Deep Contextual Schema Link Graph by Jihyung Lee, Jin-Seop Lee, Jaehoon Lee, YunSeok Choi, Jee-Hyong Lee (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-19956.txt.
Description
Jihyung Lee and four co-authors improve text-to-SQL in twenty twenty-five. The task turns a natural language question into a database query. Language models with in-context learning advanced it, but with a puzzle. Chosen demonstrations barely beat random ones, and small models lose much ground. That suggests methods lean on the raw power of huge models instead of retrieving well.
Their approach builds a Deep Contextual Schema Link Graph. It stores key information and semantic links between a question and schema items. That graph represents past examples richly, so retrieval finds truly useful demonstrations. Better examples then guide query generation. Readers get a retrieval-first design that helps smaller models punch above their weight.
Catalog record
| CREATOR | Jihyung Lee, Jin-Seop Lee, Jaehoon Lee, YunSeok Choi, Jee-Hyong Lee |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-19956