Research map / Large language models / Retrieval and knowledge
QATCH: Benchmarking SQL-centric tasks with Table Representation Learning Models on Your Data
Simone Papicchio, Paolo Papotti, Luca Cagliero
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 |
|---|---|---|---|
| Evaluate TRL models on proprietary tables instead of Spider | TAPAS/TAPEX/OmniTab/ChatGPT · QA-TRL | average metric: TAPAS 0.51 -> 0.35 (Spider -> E-commerce) | Hurt |
| Evaluate ChatGPT 3.5 on proprietary E-commerce instead of Spider | ChatGPT 3.5 · QA-TRL | average metric: 0.76 -> 0.28 | Hurt |
| Use ChatGPT 3.5 instead of TAPAS on Finance tables | ChatGPT 3.5 · QA-TRL | average metric: 0.35 -> 0.46 | Improved |
| Use ChatGPT 3.5 instead of TAPAS on E-commerce tables | ChatGPT 3.5 · QA-TRL | average metric: 0.35 -> 0.28 | Hurt |
| Switch from QA-TRL to SP-TRL (SQL generation) for ChatGPT 3.5 on E-commerce | ChatGPT 3.5 · E-commerce | average metric: 0.28 -> 0.98 | Improved |
| Evaluate ChatGPT 3.5 SP-TRL on proprietary Medicine instead of Spider dev | ChatGPT 3.5 · SP-TRL | average metric: 0.94 -> 0.99 | No clear effect |
| Evaluate GAP on proprietary Medicine instead of Spider dev | GAP · SP-TRL | average metric: 0.95 -> 0.70 | Hurt |
| Evaluate UnifiedSKG on proprietary E-commerce instead of Spider dev | UnifiedSKG · SP-TRL | average metric: 0.85 -> 0.76 | Hurt |
| Evaluate ChatGPT 3.5 SP-TRL on Spider train instead of Spider dev | ChatGPT 3.5 · SP-TRL | average metric: 0.94 -> 0.92 | No clear effect |
Closest papers by meaning
- TableBench: A Comprehensive and Complex Benchmark for Table Question Answering (AAAI 2025 · similarity 0.87)
- Text2Analysis: A Benchmark of Table Question Answering with Advanced Data Analysis and Unclear Queries (AAAI 2024 · similarity 0.87)
- ROUTE: Robust Multitask Tuning and Collaboration for Text-to-SQL (ICLR 2025 · similarity 0.87)
- LogicCat: A Chain-of-Thought Text-to-SQL Benchmark for Complex Reasoning (AAAI 2026 · similarity 0.86)
- OpenTab: Advancing Large Language Models as Open-domain Table Reasoners (ICLR 2024 · similarity 0.86)
- SQL-R1: Training Natural Language to SQL Reasoning Model By Reinforcement Learning (NeurIPS 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.
