Design of a Machine Learning Model for Automatic Generation of Domain-Specific Ontologies

Research focus

The research addresses the manual effort involved in creating, validating and updating domain ontologies. It proposes machine-learning algorithms to automate ontology generation and continuous maintenance, reducing the cost and time required for domain knowledge engineering.

From automated ontology generation to agentic knowledge operations

Agentic AI can build on this approach by orchestrating the complete knowledge lifecycle: discover source data, extract entities and relationships, generate candidate concepts, test consistency, request human validation and publish approved knowledge to downstream AI applications.

Agentic AI perspective

Modern insurance intelligence needs more than isolated models. Agentic AI connects domain knowledge, reasoning, tools, data and human approvals so that AI can move from prediction to governed execution.

UnderstandGround decisions in domain ontology, customer context and trusted enterprise data.

ReasonCombine models, rules and knowledge graphs to evaluate evidence and intent.

ActCoordinate next-best actions, updates and handoffs with traceability and human control.

What this means for insurers

Build reusable domain intelligence instead of isolated AI use cases.
Connect structured and unstructured data to contextual decisioning.
Continuously improve knowledge and recommendations with governed feedback loops.
Keep humans in the loop for high-impact decisions and exceptions.

← Back to White Papers