Multi-hop Reasoning and Retrieval in Embedding Space: Leveraging Large Language Models with Knowledge

Researchers propose EMBRAG, a new framework that combines large language models with knowledge graphs to improve reasoning accuracy and reduce hallucinations. The system generates multiple logical rules from queries and applies them in embedding space, achieving state-of-the-art performance on knowledge graph question-answering benchmarks.

• →EMBRAG framework addresses LLM hallucination issues by integrating knowledge graph retrieval for more reliable reasoning.
• →The system generates multiple logical rules from input queries and applies them in embedding space guided by knowledge graphs.
• →A reranker model interprets the generated rules and refines results for improved accuracy.
• →The approach handles multiple query interpretations and mitigates knowledge graph incompleteness and noise.
• →Experimental results show new state-of-the-art performance on two benchmark KGQA datasets.