RManLuo/Awesome-LLM-KG

Organizes research across three integration paradigms: KG-enhanced LLMs (using structured knowledge for improved reasoning and factuality), LLM-augmented KGs (leveraging language models for embedding, completion, and QA tasks), and synergized approaches combining both for knowledge representation and multi-hop reasoning. Curates papers on emerging techniques like graph foundation models for retrieval-augmented generation, temporal KG reasoning, and faithful interpretable reasoning through graph-constrained decoding. Includes application domains spanning recommendation systems, fault analysis, and knowledge base construction with links to associated code implementations.

2,572 stars. No commits in the last 6 months.