Understand Semantic Technology
Plain-English explanations of knowledge graphs, ontologies, RDF, SPARQL, OWL, Graph RAG and more, written by the consultants who build them for clients.
Knowledge Graph
A knowledge graph stores information about real things, such as people, products, places and events, together with the relationships between them. Instead of rows in separate tables, the data is a network of nodes and edges, so you can follow a chain of relationships across sources and ask questions that no single system could answer on its own.
what is a knowledge graphDigital Twin
A digital twin is a virtual model of a physical asset, process or system that is kept up to date with data from the real one. It brings sensor feeds, engineering models and context about the asset together in a structured graph, so teams can test scenarios, spot failures early and tune performance without touching the physical equipment.
what is a digital twinOntology
In information science, an ontology is a formal description of the concepts in a domain: the types of things (classes), their properties and the relationships between them. The definition most often quoted calls it a formal, explicit specification of a shared conceptualisation. Ontologies give knowledge graphs their structure and let software reason about the data.
what is an ontologyGraph RAG
Graph RAG (graph-based retrieval-augmented generation) grounds the answers of a large language model (LLM) in facts from a knowledge graph. Where standard RAG retrieves passages of text that look similar to the question, Graph RAG follows relationships in the graph to fetch the specific facts the answer needs, which reduces made-up answers and lets each answer cite where its facts came from.
what is graph RAGSemantic Search
Semantic search is search that works from the meaning of a query rather than the exact words in it. It uses ontologies, knowledge graphs and language models to work out what the user is asking for, so it can return the right documents even when they never use the words that were typed.
what is semantic searchData Fabric
A data fabric is an architecture that uses metadata, knowledge graphs and semantic technology to give people one way to find and use data across an organisation. It does not move all the data into one place. It connects data where it already lives, in clouds, databases, APIs and file systems, and describes it in a shared semantic model so it can be found, governed and queried.
what is a data fabricRDF
RDF (Resource Description Framework) is a W3C standard for representing information as statements called triples. Each triple has a subject, a predicate and an object, for example London, isCapitalOf, United Kingdom. Because every entity and relationship has its own URI, RDF data is unambiguous, can be linked across sources, and can be read by machines.
what is RDFLinked Data
Linked data is a set of principles for publishing structured data on the web so that datasets can link to each other. Tim Berners-Lee set out four rules: use URIs to name things, use HTTP URIs so people can look them up, return useful information in standard formats such as RDF when someone does, and include links to other URIs so more things can be found.
what is linked dataSPARQL
SPARQL (SPARQL Protocol and RDF Query Language) is the W3C standard language for querying and updating RDF data. It plays the part SQL plays for relational databases, but it is built around graphs: a query describes a pattern of relationships, and it can match that pattern across several data sources at once.
what is SPARQLOWL (Web Ontology Language)
OWL (Web Ontology Language) is the W3C standard for writing ontologies, which are formal descriptions of the kinds of things in a domain, their properties and how they relate. Unlike a plain taxonomy, an OWL ontology can state logical rules, such as how a class is defined or what a property is allowed to link. A reasoner reads those rules and works out facts that the data implies but never states.
what is OWL ontologySHACL
SHACL (Shapes Constraint Language) is a W3C standard for validating RDF data. You describe what valid data looks like in a set of shapes, such as which properties a record must have, what type each value is and how many values are allowed, and a SHACL engine checks the data against them. An ontology says what kinds of things can exist; SHACL says what a correct record has to contain.
what is SHACLTriplestore
A triplestore is a database built to store and query RDF triples, the subject-predicate-object statements that make up a knowledge graph. Relational databases organise data in tables and document stores in JSON; a triplestore is designed for graph-shaped data, so following relationships is fast and SPARQL is its native query language.
what is a triplestoreAI Fabric
An AI fabric is an architecture that brings knowledge graphs, data pipelines, machine learning models and AI services together in one shared layer. It builds on the data fabric by adding AI functions such as embedding generation, model serving, retrieval-augmented generation and reasoning, so AI is part of the data infrastructure from the start instead of added afterwards.
what is an AI fabricKnowledge Graph Ontology
A knowledge graph ontology is the formal schema of a knowledge graph. It defines the types of entity (classes), their attributes (properties) and the relationships between them. It is what turns a collection of connected records into knowledge that can be queried consistently and reasoned over.
knowledge graph ontologyRDF vs Property Graph
RDF (Resource Description Framework) and property graphs are the two dominant data models for knowledge graphs. RDF uses a standards-based triple model with global URIs and a rich ecosystem of W3C specifications (OWL, SHACL, SPARQL). Property graphs use a more developer-friendly model with properties on both nodes and edges, queried with languages like Cypher or Gremlin. The choice depends on your priorities: interoperability and reasoning favour RDF; developer simplicity and application-specific use cases often favour property graphs.
RDF vs property graphKnowledge Graph vs Relational Database
A knowledge graph stores data as a network of entities and relationships, which suits domains where things are densely connected and you need to follow those connections. A relational database stores data in tables of rows and columns, and is built for structured, transactional work. Most organisations need both; the question is which one fits each use case.
knowledge graph vs relational databaseFIBO (Financial Industry Business Ontology)
FIBO (Financial Industry Business Ontology) is an open ontology for the financial industry, maintained by the EDM Council. It gives formal, machine-readable definitions of financial concepts, from legal entities and corporate structures to instruments, indices and regulatory terms, so institutions can share one vocabulary for data integration, regulatory reporting and knowledge graphs.
what is FIBO ontologyOSDU (Open Subsurface Data Universe)
OSDU (Open Subsurface Data Universe) is an open, cloud-native data platform for the energy industry, developed through The Open Group. It gives operators a vendor-neutral way to store, manage and share subsurface and operational data, such as well logs, seismic surveys, production records and engineering models, with common data definitions so the data can be found and reused across teams and tools.
what is OSDUEMMO (Elementary Multiperspective Material Ontology)
EMMO (Elementary Multiperspective Material Ontology) is a top-level ontology for materials science, developed with the European Materials Modelling Council (EMMC). It gives a physics-based framework for describing materials, their properties, how they are made and how they are characterised, so simulation software, experimental databases and manufacturing systems can exchange data.
what is EMMO ontologySemantic Infrastructure
Semantic infrastructure is the set of standards, technologies and practices that make an organisation’s data understandable to machines. It includes ontologies, knowledge graphs, taxonomies, metadata management and semantic integration. With it in place, AI, analytics and applications work with data that carries its own meaning, instead of depending on hard-coded mappings that break when systems change.
what is semantic infrastructureContext Graph
A context graph is a knowledge graph used specifically to give AI systems and applications the context they need to make decisions. The term puts the emphasis on what the graph supplies: facts about how entities relate, which a language model, a recommendation engine or an agent can use to work out what a request is about instead of relying on keyword matches.
what is a context graph