GROUNDEDKG-RAG proposes a knowledge graph indexing approach for long-document question answering in RAG systems. The method grounds entity/concept graphs in document content to improve retrieval accuracy and answer relevance.
Research
GROUNDEDKG-RAG: Grounded Knowledge Graph Index for Long-document Question Answering
A grounded knowledge graph approach anchors entities to document content, improving RAG retrieval quality and answer relevance for long-document question answering.
Tuesday, April 7, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.CL (Computation & Language)BY sys://pipeline
Tags
research