Concepts

Knowledge Graphs

Data structures representing knowledge as networks of entities and relationships, enabling reasoning, connection discovery, and semantic queries over complex domains.

seed#knowledge-graph#graph#ontology#relationships#semantic-web#linked-data

What it is

A knowledge graph is a structure that represents information as a network of entities (nodes) connected by relationships (edges). Unlike a tabular database, a graph naturally captures connections between concepts, enabling discovery of implicit relationships and reasoning over complex domains.

This site uses a knowledge graph: each concept is a node, and relationships between concepts are edges forming a navigable network.

Basic structure

(Entity) --[Relationship]--> (Entity)

(Git) --[is base of]--> (GitHub)
(GitHub) --[has feature]--> (GitHub Actions)
(GitHub Actions) --[implements]--> (CI/CD)

Triples

The fundamental unit is the triple: subject → predicate → object.

(LLM) --[uses]--> (Transformer)
(RAG) --[reduces]--> (Hallucinations)
(MCP) --[standardizes]--> (Tool Use)

Types of knowledge graphs

  • Taxonomic: classification hierarchies (is-a, part-of)
  • Semantic: rich relationships with defined types (ontologies)
  • Property graphs: nodes and edges with arbitrary properties
  • RDF: W3C standard for linked data on the web

Applications with AI

GraphRAG

Combining knowledge graphs with RAG:

  • Navigate relationships to find relevant context
  • Answer questions requiring connecting multiple facts
  • Better than vector RAG for relationship questions

Grounding

Using the graph as a source of truth to reduce hallucinations.

Discovery

Finding non-obvious connections between concepts — exactly what this site's /graph page does.

Tools

ToolType
Neo4jGraph database
Amazon NeptuneManaged graph on AWS
NetworkXPython graph library
D3.js / vis.jsGraph visualization

Why it matters

Knowledge graphs represent information as entities and relationships, enabling the discovery of connections that are not evident in tabular data. They are the foundation of recommendation systems, semantic search, and knowledge bases like the one powering this site.

References

Concepts