Semantic knowledge graph
A semantic knowledge graph (SKG) is a structured knowledge representation framework that models real-world entities, concepts, events, and their relationships using graph-based data structures enriched with semantic meaning. Unlike traditional relational databases or hyperlink-based information systems, a semantic knowledge graph represents knowledge as interconnected nodes and edges that are formally defined through ontologies, enabling both humans and machines to understand the meaning of the relationships between data. Semantic knowledge graphs integrate principles from knowledge representation, graph theory, the Semantic Web, artificial intelligence, and linked data to support advanced information retrieval, reasoning, and decision-making.
A semantic knowledge graph typically represents knowledge as entities connected through explicitly defined relationships while incorporating standardized vocabularies, taxonomies, and ontologies that describe the meaning of those relationships. The approach enables information originating from multiple heterogeneous sources to be integrated into a unified, machine-readable knowledge network that can support semantic search, question answering, recommendation systems, data integration, and automated reasoning. Semantic knowledge graphs have become an important component of modern search engines, enterprise knowledge management platforms, scientific databases, digital libraries, financial intelligence systems, and Large language model (LLM) applications.
History
The conceptual foundations of semantic knowledge graphs originated in the fields of artificial intelligence, knowledge representation, and semantic networks during the 1960s and 1970s. Early semantic networks represented concepts as interconnected nodes linked by labeled relationships, providing a graphical method for representing knowledge and inference.
During the 1990s and early 2000s, the development of the World Wide Web Consortium (W3C) Semantic Web initiative expanded these concepts through standardized technologies including the Resource Description Framework (RDF), the Web Ontology Language (OWL), and the SPARQL Protocol and RDF Query Language (SPARQL). These standards enabled knowledge from diverse sources to be represented in interoperable, machine-readable formats.
The term knowledge graph gained widespread recognition after the introduction of the Google Knowledge Graph in 2012, although similar graph-based knowledge systems had existed in academic research and enterprise information management for many years. Since then, semantic knowledge graphs have been widely adopted across government, scientific research, finance, healthcare, cybersecurity, and enterprise information systems.
Architecture
A semantic knowledge graph generally consists of several interconnected components:
- Entities representing real-world objects or concepts.
- Relationships connecting entities through semantically meaningful links.
- Classes that categorize entities into hierarchical groups.
- Ontologies defining concepts, relationships, constraints, and logical rules.
- Properties describing characteristics or attributes of entities.
- Identifiers using globally unique Uniform Resource Identifiers (URIs).
- Inference mechanisms that derive new knowledge from existing facts.
Knowledge is commonly represented as a graph in which nodes correspond to entities or concepts and edges represent relationships between them.
Knowledge representation
Most semantic knowledge graphs are based on the Resource Description Framework (RDF), which represents knowledge using triples consisting of:
- Subject
- Predicate
- Object
For example:
- Microsoft — headquartered in → Redmond
- World Health Organization — specialized agency of → United Nations
- Apple Inc. — develops → iPhone