KG-RAG4SM boosts schema matching with knowledge-graph RAG, +39.79% precision over LLM SOTA
Chuangtao Ma · hf · 2026-11-10
Researchers from Aalborg University and others introduce KG-RAG4SM, a knowledge-graph-based retrieval-augmented generation method for schema matching that reduces LLM hallucinations without retraining. It combines LLM-based, vector-based, graph-traversal and hybrid retrieval with advanced reranking to pull relevant subgraphs from external KGs. It improves precision by 39.79% and F1 by 30.56% over LLM-based SOTA (e.g., Jellyfish-8B) on MIMIC, and precision by 69.20% and F1 by 21.99% over PLM-based SOTA (e.g., SMAT) on Synthea, scaling from commonsense to domain-specific KGs.
More from Research
- Stanford AI100 releases third report: 'AI Goes Mainstream' on the scaling era — StanfordHAI · 2026-11-11
- 26,040 agent runs and 100B tokens: FrontierHarness v1.1 ranks harnesses across 5 models — stuffyokodraws · 2026-11-11
- Perceptron open-sources EgoHome, a sub-second egocentric household-video benchmark — lukas_m_ziegler · 2026-11-11
- Epoch AI chart: AI cost falling ~13x a year, faster than any tech in history — sebkrier · 2026-11-11
- AI Trials Observatory launches: 93 RCTs on how AI affects mental health and beliefs — DG_Rand · 2026-11-11
- Raschka: open LLMs have abandoned full attention — hybrids and linear models now split 50/50 — rasbt · 2026-11-11