TL;DR: Knowledge Graphs and LLMs

Knowledge graphs provide structured representations of information that can enhance the reasoning capabilities of large language models. By explicitly modeling concepts and relationships, KGs offer a complementary approach to the statistical knowledge learned by LLMs, enabling more systematic and interpretable AI systems.

There are several key manners of having LLMs work with Knowledge Graphs

• Constructing graphs
• Querying graphs
• Predicting and classifying things about graph
• Navigating graphs
• Multi-hop reasoning and Retrieval Augmented Generation

erDiagram
    ENTITY_A {
        string attributes
    }
    ENTITY_B {
        string attributes
    }
    ENTITY_C {
        string attributes
    }
    ENTITY_A ||--o{ ENTITY_B : "Relationship"
    ENTITY_B ||--o{ ENTITY_C : "Relationship"
    ENTITY_C ||--o{ ENTITY_A : "Relationship"

Background