ReMatch method improves schema matching using retrieval-enhanced LLMs without the need for predefined mapping or model training.
ReMatch introduces a novel method for schema matching using retrieval-enhanced Large Language Models (LLMs), eliminating the need for predefined mapping, model training, or access to source data. The approach significantly improves matching capabilities and outperforms other machine learning methods.