Research map / Large language models / Retrieval and knowledge
AutoLink: Autonomous Schema Exploration and Expansion for Scalable Schema Linking in Text-to-SQL at Scale
Ziyang Wang, Yuanlei Zheng, Zhenbiao Cao, Xiaojin Zhang, Zhongyu Wei, Pei Fu, Zhenbo Luo, Wei Chen, Xiang Bai
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add retrieve_schema action (ablation) | AutoLink · Spider 2.0-Lite (initial top-n=100) | SRR: 84.5 -> 91.2 | Improved |
| add explore_schema action (ablation) | AutoLink · Spider 2.0-Lite (initial top-n=100) | SRR: 88.8 -> 91.2 | Improved |
| add verify_schema action (ablation) | AutoLink · Spider 2.0-Lite (initial top-n=100) | SRR: 89.6 -> 91.2 | Improved |
| add retrieve_schema action (ablation) | AutoLink · Spider 2.0-Lite (initial top-n=5) | SRR: 72.4 -> 79.2 | Improved |
| agentic iterative linking instead of full-schema input (MCS-SQL, SQL-to-Schema, RSL-SQL) | schema linking · Spider 2.0-Lite | SRR: 64.0 -> 91.2 | Improved |
| increase sampling decoding iterations of full-schema methods | MCS-SQL, SQL-to-Schema · Spider 2.0-Lite | SRR | No clear effect |
| increase initial top-n for schema retrieval (ablation) | AutoLink · Spider 2.0-Lite | EX | Mixed |
| use AutoLink on databases with over 3,000 columns | AutoLink vs baselines · Spider 2.0-Lite (>3000 columns) | SRR: baselines below 40%, AutoLink near 90% | Improved |
| swap Gemini-1.5-Pro for DeepSeek-v3 as SQL generator (ablation) | AutoLink · Bird Dev | EX: 68.71 -> 66.36 | Hurt |
Closest papers by meaning
- Structure-Guided Large Language Models for Text-to-SQL Generation (ICML 2025 · similarity 0.88)
- Spider 2.0: Evaluating Language Models on Real-World Enterprise Text-to-SQL Workflows (ICLR 2025 · similarity 0.87)
- Alpha-SQL: Zero-Shot Text-to-SQL using Monte Carlo Tree Search (ICML 2025 · similarity 0.87)
- ROUTE: Robust Multitask Tuning and Collaboration for Text-to-SQL (ICLR 2025 · similarity 0.86)
- SQLFixAgent: Towards Semantic-Accurate Text-to-SQL Parsing via Consistency-Enhanced Multi-Agent Collaboration (AAAI 2025 · similarity 0.85)
- CHASE-SQL: Multi-Path Reasoning and Preference Optimized Candidate Selection in Text-to-SQL (ICLR 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
