Semantic models, or ontologies, are key to building powerful knowledge graphs that help organizations drive explainable and trustworthy business decisions. They enrich existing data with context and meaning that humans and machines can interpret so that domain-specific knowledge can be utilized across the organization or by machines and AI applications. Semantic models:

• Bring meaning to data through the formalization of precise definitions of data and logical connections between entities
• Standardize and classify data by tying-in hierarchical vocabularies and taxonomies
• Enrich data with context by connecting additional metadata, provenance information, and governance rules

Without a semantic model, a knowledge graph is simply a graph structure that visualizes interlinked data. The added contextual richness promotes comprehensive analysis and knowledge discovery, creation and sharing that was previously unachievable. It provides the real, curated knowledge behind symbolic AI solutions that can be used to complement data-driven AI solutions with a comprehensive layer of trust, explainability and precision to Machine Learning and Large Language Models.

Topics covered

In this talk, we will introduce the gold standard for semantic knowledge modeling based on metaphactory's visual and user-friendly interface. We will discuss why it is crucial to actively involve SMEs and business users in the modeling process and demonstrate how metaphactory enables both technical and non-technical users to explicitly capture domain knowledge and expertise.

Additionally, we'll discuss how metaphactory supports the management of additional assets such as hierarchical vocabularies and taxonomies, data catalogs, and instance data, and the tie-in of semantic models with these assets to enable:

• Alignment within and between vocabularies by utilizing relation definitions from the ontology to capture more complex and multiple hierarchies, e.g., part of and functional relations in the same vocabulary between terms
• Curation of instance data and metadata, driven and guided by the underlying ontology via semantic forms, a visual canvas, or a combination thereof, incl. support for clear governance and guardrails to allow for simple and easy data curation in the form of annotations, metadata curation, etc.
• Building ontology-driven search, discovery and visualization interfaces that utilize relations and properties in the graph for powering facets, visually formulating queries or guiding users along paths in the graph

Finally, we'll cover some best practices for building semantic knowledge models and combining the power of symbolic AI and data-driven AI to power use cases such as enterprise architecture modeling, product lifecycle management, or business knowledge modeling. We'll discuss how our approach has helped customers democratize domain knowledge that was previously hidden in experts' minds, documents, or hard-coded into applications and allow this knowledge to actively drive explainable and trustworthy business decisions.