Glossary/Knowledge Representation

Ontology

An Ontology is a formal specification of concepts, categories, relationships, and rules that define and organize knowledge within a domain, enabling machines to understand meaning and relationships.

An ontology is essentially a structured vocabulary and set of rules for a domain. It defines what types of entities exist (Customer, Product, Order, Invoice), what properties each can have (Customers have names, addresses, purchase history), what relationships are possible (Customer places Order, Order contains Product), and what rules govern them (total_order_amount = sum of line item amounts). Ontologies move beyond simple schema definitions to include semantic meaning: they declare that both "buyer" and "purchaser" refer to the same concept.

Ontologies enable multiple applications to understand concepts consistently. Without an ontology, one system might define "customer" to include prospects, while another excludes them. With a shared ontology, integration and querying across systems becomes reliable. Ontologies also enable inference: if the ontology says "employees are people" and "people have birth dates," then the system can infer that employees have birth dates without explicitly stating it.

Ontologies range from simple (informal taxonomies) to complex (formal logic specifications in OWL or RDF). Simple ontologies are more practical and widely deployed, especially in enterprise data governance. Formal ontologies are used in research, knowledge representation, and AI systems where reasoning is critical. Building ontologies is labor-intensive, requiring domain experts to encode their knowledge precisely.

Key Characteristics

  • Defines entity types, properties, and relationships within a domain
  • Specifies constraints and rules governing valid combinations
  • Enables inference: deriving conclusions from stated facts and rules
  • Supports multiple inheritance and complex relationship types
  • Provides semantic meaning beyond syntactic schema definitions
  • Can range from informal controlled vocabularies to formally-specified logic systems

Why It Matters

  • Enables consistent understanding of concepts across teams and systems
  • Facilitates data integration by mapping concepts between sources
  • Supports inference and reasoning enabling more sophisticated analytics
  • Provides grounding for AI systems, improving accuracy and reducing hallucination
  • Enables knowledge discovery by making relationships and rules explicit
  • Supports compliance and governance by defining authorized concepts and relationships

Example

An e-commerce ontology defines: Product (sku, name, price), Customer (name, address, segment), Order (customer, products, total), and rules: "total order amount must equal sum of product prices for products in order; customer segment determines eligible discounts." This enables both validation and reasoning about customer behavior.

Coginiti Perspective

In Coginiti, SMDL (Semantic Model Definition Language) functions as an ontology for analytics, defining entities, dimensions, measures, and relationships that formalize domain knowledge. Through SMDL's relationships (one_to_one, one_to_many, many_to_one), the semantic layer enforces business logic and enables consistent interpretation across tools, while Semantic SQL's implicit joins automatically apply these ontological rules during query execution.

See Semantic Intelligence in Action

Coginiti operationalizes business meaning across your entire data estate.