Human-Level Text-to-SQL via RLVR and Verified Data Cleaning
ddkang · x · 2026-08-28
Paper proposes achieving human-level Text-to-SQL by fine-tuning Qwen3-235B using RLVR (Reinforcement Learning on Verified Data) on clean data, without complex pipelines.
Key findings:
- Existing training data contains pervasive annotation errors that mislead optimization.
- Developed a multi-round expert verification pipeline to curate BIRD-Platinum (2.5k instances from BIRD Train), correcting errors in 61% of instances.
- Fine-tuning on BIRD-Platinum yields 11-16% improvements over BIRD Train on Arcwise-Plat and Spider2, outperforming SOTA open-source systems.
More from Research
- Schmidt Sciences awards $10M to 19 AI humanities research teams — OdedRechavi · 2026-10-15
- Mistral and CERN partner to build a Large Physics Model trained on billions of simulated collisions — MistralAI · 2026-10-15
- Global Transport coupling lifts guided flow generation: ImageNet FID 2.06 to 1.91 — mmbronstein · 2026-10-15
- Your breathing gives you away: Wi-Fi signals can reveal your breath rhythm — TinfoilTricorn · 2026-10-15
- Beyond text correlation: attaching a geometry simulator to ground model reasoning — gerardsans · 2026-10-15
- Sequen CEO Zoe Weil: Ranking Models Should Optimize Purchases, Not Clicks — TWIML AI Podcast · 2026-10-15