Semantic Graphs - An Introduction

A semantic graph is both a graph model and a knowledge representation. It represents knowledge using nodes and edges while also capturing the meaning of that knowledge in a structured form that can be used for reasoning and inference. In a semantic graph, nodes represent entities or concepts, while edges represent relationships between them.

Semantic Models

A semantic graph and semantic model should not be confused as they have very different meanings despite being related concepts. These terms can get confusing since a semantic graph is a graph model. 

A semantic model is a broader concept than a semantic graph. The term "semantic model" refers to a conceptual framework or a set of formal rules and definitions that define the meaning and interpretation of data within a specific domain. 

Semantic models are often represented as Ontologies in a model called OWL. They specify how entities, attributes, and relationships are defined and organised, along with the rules governing their behaviour and constraints.

A semantic model which is an Ontology typically provides a structured and standardised representation of information, allowing for consistent interpretation and understanding of data across different systems and applications. It serves as a foundation for data integration, interoperability, and semantic consistency.


An ontology is a formal and explicit specification of a conceptualization of a domain. In simpler terms, it is a structured and organised representation of knowledge about a specific domain of interest. Ontologies are commonly used in various fields, including computer science, artificial intelligence, information science, and philosophy.

In the context of computer science and artificial intelligence, ontologies play a crucial role in knowledge representation. They help to model the relationships and properties of entities within a domain, enabling machines to reason, understand, and make inferences about the domain.

Key components of an ontology include:

  • Concepts or Classes: These represent the entities or things within the domain. Each concept represents a category of objects with common attributes and behaviours.
  • Properties or Relations: These define the relationships between concepts. For example, "hasParent" can be a relation between the concept "Child" and "Parent."
  • Instances or Individuals: These are specific objects or examples belonging to a particular concept. For example, "John" and "Mary" could be instances of the concept "Person."
  • Axioms or Constraints: These are logical statements that impose rules and constraints on the concepts and their relationships, helping to ensure the consistency and correctness of the ontology.

Ontologies are typically represented in a formal language, such as Web Ontology Language (OWL) for the Semantic Web or Resource Description Framework (RDF) for linked data. These languages provide a standardised way to define and reason about ontologies.

Ontologies have various applications, including but not limited to:

  • Semantic web technologies, where they enhance data integration and interoperability on the web.
  • Knowledge engineering, facilitating the development of intelligent systems and expert systems.
  • Natural language processing and information retrieval, enabling better understanding of natural language by machines.
  • Domain-specific modelling and data organisation in various industries and research domains.

Overall, ontologies provide a powerful tool for organising, sharing, and reasoning about knowledge in a structured and standardised manner, making them invaluable for a wide range of applications in modern computing and knowledge representation.

Semantic Graphs vs Semantic Networks

In general, the terms "semantic graph" and "semantic network" are used interchangeably to refer to a type of knowledge representation that captures knowledge as a network of interconnected entities and their relationships.

Both semantic graphs and semantic networks use nodes to represent entities and edges to represent relationships between those entities. These relationships are often labelled with semantic information to provide meaning and context to the data.

However, some researchers and practitioners may use these terms differently depending on their specific context or application. For instance, some may use "semantic network" to refer to a specific type of graph model that has a more restricted set of relationship types, while others may use "semantic graph" to refer to a more general type of graph that can represent a wider range of relationships.

In general, the key idea is that both semantic graphs and semantic networks are graph-based knowledge representations used to capture semantic relationships between entities.

Semantic Graph Use Cases

Semantic graphs are used in various applications, including natural language processing, machine learning, and knowledge management. They allow for more efficient and accurate information processing by representing it in a structured way that can be easily analysed and queried. In the following sections, we will cover some fundamental concepts that are important to understand in the field of semantic graphs. If you have yet to see it, our explainer about ontologies is worth reading.

RDF and RDF Graphs

RDF stands for Resource Description Framework. It is a standard format for representing semantic graphs and semantic models; often referred to as Ontologies. RDF provides a way to describe resources such as web pages, books, or people in a machine-readable format.

RDF is based on the idea of subject-predicate-object triples, also known as RDF triples. An RDF triple consists of a subject, a predicate, and an object. The subject is a resource that is being described, the predicate is a property or relationship of the resource, and the object is the value of that property or relationship.

RDF is often used to represent data in a graph format, where resources (such as web pages, people, or events) are nodes in the graph and predicates (such as "is a member of" or "was created by") are edges that connect the nodes. This makes it easy to represent complex relationships between resources and to query the data using tools such as SPARQL.

RDF is a key technology in the Semantic Web, which aims to make web data machine-readable and interoperable across different applications and platforms.

RDF Graph Use Cases

RDF graphs are particularly suited to representing data that involves complex relationships and ontologies. Some examples of data that can be represented using RDF graphs include:

  1. Scientific data: RDF graphs can be used to represent scientific data and metadata, including information about research publications, datasets, and experimental results.
  2. Cultural heritage data: RDF graphs can be used to represent data about cultural heritage objects, including information about artworks, archaeological finds, and historic buildings.
  3. Linked open data: RDF graphs can be used to represent and link data from different sources on the web, allowing users to query and analyse data from a variety of different domains.
  4. Government data: RDF graphs can be used to represent government data, including information about policies, legislation, and public services.
  5. Healthcare data: RDF graphs can be used to represent healthcare data, including information about patients, medical procedures, and clinical trials.

These are just a few examples of the kinds of data that can be represented using RDF graphs. The flexibility of RDF means that it can be used to represent many different kinds of data, from social network data to e-commerce data to environmental data.

RDF and Ontologies

Ontologies are typically represented in RDF syntax.

Ontologies typically use RDF to represent the relationships between concepts, but they also include additional information, such as axioms and rules that specify the semantics of the ontology. Some popular ontology languages that use RDF include OWL and RDFS.

Ontology Languages

Ontology languages are used to represent and reason about knowledge. Each language has its own strengths and weaknesses and is best suited for use in different domains and applications.

OWL (Web Ontology Language)

OWL is an ontology language that uses RDF as one of its underlying technologies. Like RDF, OWL is a formal language for representing knowledge, but it provides more expressive power and logical formalisms for creating more complex ontologies.

OWL is based on the same underlying concepts as RDF, such as the representation of information as statements with subjects, predicates, and objects. However, OWL adds additional constructs that allow for more precise modelling of complex relationships between concepts. For example, OWL supports the representation of classes, properties, and individuals, as well as more sophisticated constructs such as transitive properties, disjointness, and cardinality restrictions.

RDFS (RDF Schema)

RDFS is a language that extends RDF with vocabulary for defining classes, properties, and subsumption relationships between them. RDFS provides a simple and lightweight ontology language but has less expressive power than OWL.

Common Logic (CL)

Common Logic is a family of logical languages designed for representing and reasoning about knowledge in a wide range of domains. Common Logic supports many logical constructs, making it more expressive than many other ontology languages.


CycL is a knowledge representation language designed to represent common sense knowledge. CycL includes a rich set of axioms and rules for reasoning about various concepts and relationships.

Description Logics (DL)

Description Logics are a family of formal languages used for representing and reasoning about knowledge. DLs are based on first-order logic and have well-defined semantics, making them useful for applications such as ontology engineering and automated reasoning.

Types of Semantic Graphs

The type of semantic graph is not determined by the chosen ontology language alone but rather by the combination of the ontology language and the specific knowledge representation task at hand.

Different ontology languages provide different levels of expressive power and formalism for representing knowledge. Still, how the knowledge is represented as a semantic graph depends on the specific use case and the modelled domain.

For example, RDF is a simple and flexible data model for representing information and can be used to represent a wide range of knowledge. However, when modelling more complex domains, an ontology language such as OWL may be necessary to provide the required level of formalism and expressiveness. In this case, the resulting semantic graph would be more complex and expressive than an RDF-based semantic graph. Here are some examples of different types of semantic graphs:

Simple RDF Graph

A simple RDF graph is a graph-based data structure representing information as subject-predicate-object triples. This type of graph is commonly used to represent basic knowledge about entities and their relationships.

Complex OWL Ontology Graph

An OWL ontology graph is a more complex semantic graph representing knowledge using OWL constructs such as classes, properties, and individuals. This type of graph can represent more complex relationships between concepts and is commonly used in applications such as biomedical informatics, where complex taxonomies of diseases and treatments need to be represented.

Rule-Based Semantic Graph

A rule-based semantic graph represents knowledge using a set of logical rules and constraints. This type of graph can be used to reason about complex relationships between entities and infer new knowledge based on the rules.

Semantic Web Technologies

The semantic web refers to a set of technologies and standards designed to enhance the meaning and structure of information on the World Wide Web. It aims to enable machines to understand and process web content in a more intelligent and automated manner. At the heart of the semantic web are semantic graph technologies, which play a crucial role in representing and organising the underlying data.

Semantic web technologies and semantic graphs are closely intertwined. Semantic graphs serve as the underlying data representation mechanism, capturing the relationships and connections between entities. They enable interoperability, knowledge discovery, and automated reasoning on the semantic web. Semantic web technologies, such as the Resource Description Framework (RDF) and the Web Ontology Language (OWL), provide the means to define and represent semantic graphs.

By leveraging the structured and semantically rich information captured in semantic graphs, applications can provide personalised recommendations, perform advanced search and discovery, support knowledge-based reasoning, and facilitate data integration across various domains. All of this supports the development of intelligent applications and services.


How to Create a Semantic Graph

The Graph.Build Studio Allows you to very easily build a semantic graph model with no code. You can import data from virtually any source using the enterprise-ready transformers and publish your model to an output file or directly to a database of your choosing.